docket no: a-93-02 v-b-14 2019/us epa... · docket no: a-93-02 v-b-14 technical support document...

DOCKET NO: A-93-02V-B-14

TECHNICAL SUPPORT DOCUMENT FOR

SECTION 194.23: PARAMETER JUSTIFICATION REPORT

U. S. ENVIRONMENTAL PROTECTION AGENCYOffice of Radiation and Indoor Air

Center for the Waste Isolation Pilot Plant401 M. Street, S. W.

Washington, DC 20460

MAY 1998

Technical Support for this document was provided by TechLaw, Inc., and its subcontractorsunder EPA Contract 68D50174.

ES - 3

LIST OF ACRONYMS

BIR Baseline Inventory ReportCCA Compliance Certification ApplicationCCDF Complementary Cumulative Distribution FunctionCMS Configuration Management SystemDBR Direct Brine ReleaseDOE U.S. Department of EnergyDRZ Disturbed Rock ZoneEEP Expert Elicitation PanelEPA U.S. Environmental Protection AgencyJR Parameter Justification ReportPA Performance AssessmentPAVT Performance Assessment Verification TestPR Parameter ReportQA Quality AssuranceSA Sensitivity AnalysisSNL Sandia National LaboratoriesTDEM Time Domain ElectromagneticWIPP Waste Isolation Pilot Project

ES - 1

EXECUTIVE SUMMARY

This report describes the disposition of 58 Waste Isolation Pilot Project (WIPP) performanceassessment (PA) input parameters that were found by the Agency to be inadequately supported.Each of these parameters was included in the database for the U.S. Department of Energy’s(DOE) WIPP Compliance Certification Application (CCA) of October 1996. The finding ofinadequate support was made by the Agency following a comprehensive review of nearly 1,600parameters used in the WIPP PA. The inadequately supported parameters were identified toDOE in Enclosures 2, 3, and 4 of an Agency letter dated March 19, 1997, and many of thoseparameters were included in a sensitivity analysis performed by the Agency. The inadequatelysupported parameters were so identified because they were potentially important to the results ofthe PA and they lacked supporting data, they had different values or ranges than were supportedin the DOE database, or they had questionable values or ranges.

Each of the inadequately supported parameters was further reviewed by the Agency and waseither resolved and found to no longer be in question, or was included in the EPA-mandatedPerformance Assessment Verification Test (PAVT). A summary of the inadequately supportedparameters and their disposition is presented in Tables ES-1 through ES-3. The mandated PAVTrepeats the performance assessment calculations presented by DOE in the CCA but withmodifications to the codes to correct errors identified by DOE and the Agency, and with revisedvalues, ranges, and distributions determined by the Agency for those parameters that remained inquestion. Those parameters found to be no longer in question were typically either found to benot sensitive in the Agency’s sensitivity analysis, the parameter was accepted after review ofadditional documentation provided by DOE or through the Agency’s independent analysis, theparameter was eliminated because of a change in the model, or the parameter was found to nothave been used in the CCA version of the PA model. In the last case, if the parameter or its usewas considered by the Agency to be potentially important, alternate parameters were identifiedand changed in the PAVT to achieve the Agency’s objective in questioning the originalparameter.

After making the necessary adjustments to allow for model changes, a final list of 22 parametersthat were to be changed in the PAVT was developed and is presented in Table ES-4. This tablealso summarizes the parameter values, ranges, and distributions used in the PAVT as well as theoriginal values used in the CCA. The basis for selecting each parameter, value, range, anddistribution is presented in Sections 3, 4, and 5 of this report.

ES - 2

Table ES-1. Parameters Lacking Supporting DataNo. ID

No.Material ID -Parameter ID

Description Disposition Reason

1 3245 BLOWOUT-CEMENT

Waste cementationstrength

Change required forPAVT in 4/25 letterbut removed

Sensitive but removeddue to change in model

2 3246 BLOWOUT-PARTDIA

Waste particle diameter Required revisionusing expert elicitationprocess

Not appropriatelyjustified

3 198 DRZ_1-PRMX_LOG

Intrinsic permeability inX-direction in disturbedrock zone

Change required forPAVT in 4/17 letter

Sensitive and notappropriately justified

4 2177 S_MB_139-DPHIMAX

Incremental increase inanhydrite porosity inMarker Bed 139

Removed in 4/25 letter Documentation providedand accepted afterreview

5 2180 S_MB_139-PF_DELTA

Incremental pressure forfull fracture development


6 586 S_MB_139-PI_DELTA

Fracture initiationpressure increment


7 2178 S_MB_139-KMAXLOG

Maximum permeabilityin altered anhydrite


8 3134 BH_OPEN-PRMX_LOG

Intrinsic permeability inX-direction in openborehole

Removed in 4/25 letter Not sensitive

9 2158 S_ANH_AB-DPHIMAX

Incremental increase inanhydrite porosity inbeds A and B


10 214 EXP_AREA-PRMX_LOG

Intrinsic permeability inX-direction inexperimental area


11 3473 BLOWOUT-THCK_CAS

Thickness of Castilebrine pocket for directbrine release


12 3456 BLOWOUT-RE_CAST

Radius of Castile brinepocket for direct brinerelease


13 3194 CASTILER-GRIDFLO

Index for selecting brinepocket volume

Removed in 4/25 letter Not used in CCA PAmodel - volume changedusing other parameters

ES - 3

ES - 4

Table ES-2. Parameters with Different Values or Ranges

No. IDNo.

Material ID -Parameter ID


1 3493 GLOBAL-PBRINE

Probability ofencountering pressurizedbrine



2 2254 BOREHOLE-TAUFAIL

Waste shear resistance Change required forPAVT in 4/25 letterand 6/6 note to docket


3 3184 BH_SAND-PRMX_LOG

Long term intrinsicborehole permeability inX-direction



4 2918 CASTILER-VOLUME

Castile brine pocketvolume

Removed in 4/25 letter Not used in CCA PAmodel - volume changedusing compressibilityand porosity adjustment

5 61 CASTILER-COMP_RCK

Castile brine pocket rockcompressibility


Not appropriatelyjustified; used to changebrine pocket volume

ES - 5

Table ES-3. Parameters with Questionable Values or Ranges

No. IDNo.



1 27 BOREHOLE-DOMEGA

Drill string angularvelocity



2 64 CASTILER-POROSITY

Castile brine pocketporosity


3 66 CASTILER-PRESSURE

Castile brine pocket porepressure


4 259 PAN_SEAL-PRMX_LOG

Intrinsic permeability ofpanel seal in X-direction


5 528 S_ANH_AB-POROSITY

Porosity of anhydrite bedsA and B


6 567 S_MB138-POROSITY

Porosity of anhydriteMarker Bed 138





8 651 WAS_AREA-ABSROUGH

Waste area absoluteroughness

Removed in 4/17 letter Not sensitive andacceptable after review ofdocumentation

9 653 WAS_AREA-COMP_RCK

Waste area rockcompressibility


10 1992 WAS_AREA-DIRNCCHW

Bulk density of ironcontainers in CH waste

Removed in 4/17 letter Sensitive, butdocumentation providedand accepted after review

11 1993 WAS_AREA-DIRNCRHW

Bulk density of ironcontainers in RH waste


12 2040 WAS_AREA-DIRNCHW

Average density of iron-based material in CHwaste


13 2041 WAS_AREA-DCELLCHW

Average density ofcellulosics in CH waste


14 2274 WAS_AREA-DCELLRHW

Average density ofcellulosics in RH waste


15 2907 STEEL-CORRMCO2

Steel corrosion rate Change required forPAVT in 4/17 letter


16 3147 CONC_PLG-POROSITY

Borehole plug porosity Removed in 4/17 letter Not sensitive

No. IDNo.



ES - 6

17 3185 CONC_PLG-PRMX_LOG

Borehole plugpermeability in X-direction



18 3256 BLOWOUT-FGE

Gravity scaling factor Change required forPAVT in 4/17 letter butremoved

Removed due to changein model

19 3259 BLOWOUT-APORO

Waste permeability inCUTTINGS_S Model

Change required forPAVT in 4/25 letter butremoved

Not used in CCA PAmodel -replaced withchanges to parameters663 WAS_AREA-PRMX_LOG and 2131REPOSIT-PRMX_LOG

20 3429 PHUMOX3-PHUMCIM

Humic colloidproportionality constant


21 3471 BLOWOUT-MAXFLOW

Maximum period ofuncontrolled boreholeflow


22 3472 BLOWOUT-MINFLOW

Minimum period ofuncontrolled boreholeflow


23 3433 PHUMOX3-PHUMSIM



24 3470 GLOWOUT-GAS_MIN

DBR cutoff gas flow rate Removed in 4/25 letter Not sensitive

25 3317 PU-PROPMIC Microbial colloidproportionality constantfor plutonium


26 3405 SOLMOD6-SOLCIM

U(VI) solubility limit inCastile brine

Removed in 6/6 note todocket

Sensitive, butdocumentation providedand accepted after review

27 3409 SOLMOD6-SOLSIM

U(VI) solubility limit inSalado brine

Removed in 6/6 note todocket

Sensitive, butdocumentation providedand accepted after review

28a 3406 SOLMOD3-SOLCIM

Oxidation state +3solubility limit in Castilebrine


Use corrected model andnew mineral phase(hydromagnesite)

28b 3402 SOLMOD3-SOLSIM

Oxidation state +3solubility limit in Saladobrine







No. IDNo.



ES - 7













33 3311 AM-PROPMIC Microbial colloidproportionality constantfor americium


34 3482 AM+3-MKD_AM

Matrix partitioncoefficient for americium+3

Change in distributionrequired for PAVT in4/25 letter

Range accepted afterreview of documentation

35 3480 PU+3-MKD_PU Matrix partitioncoefficient for plutonium+3



36 3481 PU+4-MKD_PU Matrix partitioncoefficient for plutonium+4


Range accepted after review of documentation

37 3479 U+4-MKD_U Matrix partitioncoefficient for uranium +4



38 3478 TH+4 -MKD_TH

Matrix partitioncoefficient for thorium +4

Added to list, requiredfor PAVT on 4/25 letterfootnote. Changeddistribution


39 3475 U+6-MKD_U Matrix partitioncoefficient for uranium +6



40 656 WAS_AREA-GRATMICH

Gas generation rate due tomicrobial action underhumid conditions

Removed in 4/25 letter No sensitivity

41 657 WAS_AREA-GRATMICI

Gas generation rate due tomicrobial action underinundated conditions


ES - 8

Table ES-4. Parameters Changed by EPA in Performance Assessment Verification Test

ID No. Material ID -Parameter ID

Use Distribution Minimum Maximum Median Units

198 DRZ_1 -PRMX_LOG "

PAVTCCA

UniformConstant

-19.4-15.0

-12.5-15.0

-15.95-15.0

Log m2

Log m2

3184 BH_SAND -PRMX_LOG "

PAVTCCA

UniformUniform

-16.3-14.0

-11.0-11.0

-13.65-12.5

Log m2

Log m2

3185 CONC_PLG -PRMX_LOG "

PAVTCCA

UniformConstant

-19-16.3

-17-16.3

-17.3-16.3

Log m2

Log m2

663 WAS_AREA -PRMX_LOG*"

PAVTCCA

ConstantConstant

-12.6198-12.769

-12.6198-12.769

-12.6198-12.769

Log m2

Log m2

2131 REPOSIT -PRMX_LOG*"

PAVTCCA

ConstantConstant

-12.6198-12.769

-12.6198-12.769

-12.6198-12.769

Log m2

Log m2

2907 STEEL -CORRMCO2

PAVTCCA

UniformUniform

0.00.0

3.17 E-141.59 E-14

1.58 E-147.94 E-14

m/sm/s

61 CASTILER -COMP_RCK

PAVTCCA

TriangularTriangular

2.0 E-115.0 E-12

1.0 E-101.0 E-08

4.0 E-11-1.0 E-10-

Pa-1

Pa-1

8000 CASTILER -POR_BPKT j

PAVTCCA

Triangular- -

0.1848- -

0.9240- -

0.3696-- -

Dimension-less

3493 GLOBAL -PBRINE

PAVTCCA

UniformConstant

0.010.08

0.600.08

0.3050.08

Dimension-less

27 BOREHOLE -DOMEGA

PAVTCCA

CumulativeConstant

4.207.8

23.07.8

7.87.8

Radians/secRadians/sec

2254 BOREHOLE -TAUFAIL

PAVTCCA

LoguniformUniform

0.050.05

77.010.0

2.05.025

PaPa

3482 AM+3 -MKD_AM

PAVTCCA

LoguniformUniform

0.020.02

0.500.50

0.100.26

m3/kgm3/kg

3480 PU+3 - MKD_PU PAVTCCA

LoguniformUniform

0.020.02

0.500.50

0.100.26

m3/kgm3/kg


LoguniformUniform

0.9000.900

20.020.0

4.24310.45

m3/kgm3/kg

3479 U+4 - MKD_U PAVTCCA

LoguniformUniform

0.9000.900

20.020.0

4.24310.45

m3/kgm3/kg


LoguniformUniform

3.00 E-053.00 E-05

3.00 E-023.00 E-02

9.49 E-041.50 E-02

m3/kgm3/kg

3478 TH+4 - MKD_TH PAVTCCA

LoguniformUniform

0.9000.900

20.020.0

4.24310.45

m3/kgm3/kg



ES - 9

3406 SOLMOD3 -SOLSIM

PAVTCCA

ConstantConstant

1.2 E-075.82 E-08

1.2 E-075.82 E-08

1.2 E-075.82 E-08

moles/litermoles/liter

3402 SOLMOD3 -SOLCIM

PAVTCCA

ConstantConstant

1.3 E-086.52 E-08

1.3 E-086.52 E-08

1.3 E-086.52 E-08



PAVTCCA

ConstantConstant

1.3 E-084.4 E-06

1.3 E-084.4 E-06

1.3 E-084.4 E-06



PAVTCCA

ConstantConstant

4.1 E-086.0 E-09

4.1 E-086.0 E-09

4.1 E-086.0 E-09



PAVTCCA

ConstantConstant

2.4 E-072.3 E-06

2.4 E-072.3 E-06

2.4 E-072.3 E-06



PAVTCCA

ConstantConstant

4.8 E-072.2 E-06

4.8 E-072.2 E-06

4.8 E-072.2 E-06


* These parameters replaced BLOWOUT - APORO, which is not used in a current PA code- This is the mode of the triangular distribution" Parameter similarly varied in Y- and Z-directionsj New parameter and number created by DOE for the PAVT to allow brine pocket volume to be varied

ES - 10

TABLE OF CONTENTS

EXECUTIVE SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ES-1

1.0 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Background and Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Report Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

2.0 PARAMETER SELECTION PROCESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.1 Parameter Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.2 Parameter Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

3.0 PARAMETERS LACKING SUPPORTING DATA . . . . . . . . . . . . . . . . . . . . . . . . . . 73.1 3245 BLOWOUT - CEMENT: Waste Cementation Strength . . . . . . . . . . . . . . 73.2 3246 BLOWOUT - PARTDIA: Waste Particle Diameter . . . . . . . . . . . . . . . . . 83.3 198 DRZ_1 - PRMX_LOG: DRZ Permeability . . . . . . . . . . . . . . . . . . . . . . . . 93.4 2177 S_MB_139 - DPHIMAX: Maximum Incremental Increase in Anhydrite

porosity in Marker Bed 139 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103.5 2180 S_MB_139 - PF_DELTA: Incremental Pressure for Full Fracture

Development in Marker Bed 139 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123.6 586 S_MB_139 - PI_DELTA: Fracture Initiation Pressure Increment in

Marker Bed 139 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133.7 2178 S_MB_139 - KMAXLOG: Maximum Permeability in Altered Anhydrite

in Marker Bed 139 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143.8 3134 BH_OPEN - PRMX_LOG: Open Borehole Permeability . . . . . . . . . . . . 143.9 2158 S_ANH_AB - DPHIMAX: Incremental Increase in Anhydrite Porosity in

Beds A and B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153.10 214 EXP_AREA - PRMX_LOG: Experimental Area Permeability . . . . . . . 153.11 3473 BLOWOUT - THCK_CAS: Thickness of Castile Brine Pocket for Direct

Brine Release . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163.12 3456 BLOWOUT - RE_CAST: Radius of Castile Brine Pocket for Direct Brine

Release . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163.13 2918 CASTILER - GRIDFLO: Index for Selecting Brine Pocket Volume . . . 17

4.0 PARAMETERS WITH DIFFERENT VALUES OR RANGES . . . . . . . . . . . . . . . . 184.1 3493 GLOBAL - PBRINE: Probability of Encountering Pressurized Brine . . 184.2 2254 BOREHOLE - TAUFAIL: Waste Shear Resistance . . . . . . . . . . . . . . . . 194.3 3184 BH_SAND - PRMX_LOG: Long Term Borehole Permeability . . . . . . . 214.4 2918 CASTILER - VOLUME: Castile Brine Pocket Volume . . . . . . . . . . . . . 224.5 61 CASTILER - COMP_RCK: Castile Brine Pocket Rock Compressibility 22

5.0 PARAMETERS WITH QUESTIONABLE VALUES OR RANGES . . . . . . . . . . . 25

ES - 11

5.1 27 BOREHOLE - DOMEGA: Drill String Angular Velocity . . . . . . . . . . . 255.2 64 CASTILER - POROSITY: Castile Brine Pocket Porosity . . . . . . . . . . . 255.3 66 CASTILER - PRESSURE: Castile Brine Pocket Pore Pressure . . . . . . . 265.4 259 PAN_SEAL - PRMX_LOG: Panel Seal Permeability . . . . . . . . . . . . . . 265.5 528 S_ANH_AB - POROSITY: Effective Porosity of Anhydrite Beds A

and B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275.6 567 S_MB138 - POROSITY: Effective Porosity of Anhydrite Marker

Bed 138 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275.7 588 S_MB139 - POROSITY: Effective Porosity of Anhydrite Marker

Bed 139 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285.8 651 WAS_AREA - ABSROUGH: Waste Area Absolute Roughness . . . . . . 285.9 653 WAS_AREA - COMP_RCK: Waste Area Rock Compressibility . . . . . 285.10 1992 WAS_AREA - DIRNCCHW: Bulk Density of Iron Containers for CH

Waste . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295.11 1993 WAS_AREA - DIRNCRHW: Bulk Density of Iron Containers for RH

Waste . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295.12 2040 WAS_AREA - DIRNCHW: Average Density of Iron-Based Material in

CH Waste . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305.13 2041 WAS_AREA - DCELLCHW: Average Density of Cellulosics in

CH Waste . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315.14 2274 WAS_AREA - DCELLRHW: Average Density of Cellulosics in

RH Waste . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315.15 2907 STEEL - CORRMCO2: Steel Corrosion Rate . . . . . . . . . . . . . . . . . . . . . 315.16 3147 CONC_PLG - POROSITY: Borehole Plug Porosity . . . . . . . . . . . . . . . . 335.17 3185 CONC_PLG - PRMX_LOG: Borehole Plug Permeability . . . . . . . . . . . 335.18 3256 BLOWOUT - FGE: Gravity Scaling Factor . . . . . . . . . . . . . . . . . . . . . . 345.19 3259 BLOWOUT - APORO: Waste Permeability in CUTTINGS_S Model (663

WAS_AREA - PRMX_LOG)(2131 REPOSIT - PRMX_LOG) . . . . . . . 355.20 3429 PHUMOX3 - PHUMCIM: Humic Colloid Proportionality Constant in

Castile Brine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365.21 3471 BLOWOUT - MAXFLOW: Maximum Period of Uncontrolled Borehole

Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375.22 3472 BLOWOUT - MINFLOW: Minimum Period of Uncontrolled Borehole

Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375.23 3433 PHUMOX3 - PHUMSIM: Humic Colloid Proportionality Constant in

Salado Brine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385.24 3470 BLOWOUT - GAS_MIN: DBR Cutoff Gas Flow Rate . . . . . . . . . . . . . 385.25 3317 PU - PROPMIC: Microbial Colloid Proportionality Constant for

Plutonium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395.26 3405 SOLMOD6 -SOLCIM: U(VI) Solubility in Castile Brine . . . . . . . . . . . 395.27 3409 SOLMOD6 -SOLSIM: U(VI) Solubility in Salado Brine . . . . . . . . . . . . 41

5.28 3402 SOLMOD3 - SOLCIM: Oxidation State +III Solubility in Castile Brine(3406 SOLMOD3 - SOLSIM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

ES - 12

5.29 3403 SOLMOD4 - SOLCIM: Oxidation State +IV Solubility in Castile Brine 435.30 3407 SOLMOD4 - SOLSIM: Oxidation State +IV Solubility in Salado Brine 435.31 3404 SOLMOD5 - SOLCIM: Oxidation State +V Solubility in Castile Brine 445.32 3408 SOLMOD5 - SOLSIM: Oxidation State +V Solubility in Salado Brine . 455.33 3311 AM - PROPMIC: Microbial Colloid Proportionality Constant for

Americium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 455.34 3482 AM+3 - MKD_AM: Matrix Partition Coefficient for Americium +III . . 465.35 3480 PU+3 - MKD_PU: Matrix Partition Coefficient for Plutonium +III . . . . 485.36 3481 PU+4 - MKD_PU: Matrix Partition Coefficient for Plutonium +IV . . . . 485.37 3479 U+4 - MKD_U: Matrix Partition Coefficient for Uranium +IV . . . . . . . 495.38 3475 U+6 - MKD_U: Matrix Partition Coefficient for Uranium +VI . . . . . . . 495.39 656 WAS_AREA - GRATMICH: Gas Generation Rate due to Microbial

Action under Humid Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 505.40 657 WAS_AREA - GRATMICI: Gas Generation Rate due to Microbial Action

under Inundated Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

6.0 SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

LIST OF TABLES

Table ES-1 Parameters Lacking Supporting Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ES-2Table ES-2 Parameters with Different Values or Ranges . . . . . . . . . . . . . . . . . . . . . . . . . . ES-4Table ES-3 Parameters with Questionable Values or Ranges . . . . . . . . . . . . . . . . . . . . . . . ES-5Table ES-4 Parameters Changed by EPA in Performance Assessment Verification Test . . .ES-8Table 2.1-1 Inadequately Supported Parameters Identified in EPA’s March 19, 1997 Letter . .5Table 6.1 Parameters Lacking Supporting Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53Table 6.2 Parameters with Different Values or Ranges . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55Table 6.3 Parameters with Questionable Values or Ranges . . . . . . . . . . . . . . . . . . . . . . . . . 56Table 6.4 Parameters Changed by EPA in Performance Assessment Verification Test . . . . .60

1

1.0 INTRODUCTION

EPA conducted a comprehensive review of the supporting rationale for the parameters used inthe performance assessment (PA) calculations presented by the U.S. Department of Energy(DOE) in the October 1996 Compliance Certification Application (CCA) for the Waste IsolationPilot Plant (WIPP). Those parameters found to be inadequately supported were identified by theAgency to DOE in March 1997. This report identifies those parameters, presents the basis fortheir original selection, and describes the subsequent evaluations that were performed to identifythe parameters of primary concern for regulatory compliance.

1.1 Background and Scope

This report is one of a series of three reports that provide detailed documentation of EPA’stechnical review of the CCA and the methodology used by the Agency to evaluate DOEcompliance with the requirements of 40 CFR 194.23(c)(4). These three reports are brieflydescribed in the following paragraphs.

The first report, Technical Support Document for Section 194.23 - Parameter Report (PR)[Docket A-93-02, Item V-B-12], describes EPA’s detailed review of the DOE’s supportingdocumentation and technical rationale for the parameters used in the PA model. The reportdescribes the screening process used by the Agency to identify those parameters that were poorlydocumented, that had a weak technical basis, and that may be important in determiningcompliance. This screening occurred in several steps and culminated in identifying a series ofparameters that warranted further review. Those parameters are listed Tables ES-1, ES-2, andES-3 of this report.

The second report, Technical Support Document for Section 194.23 - Sensitivity Analysis Report(SA) [Docket A-93-02, Item V-B-13], describes the Agency’s evaluation of key PA modeloutputs to changes in selected input parameters. The input parameters selected for this analysiswere based primarily on the results of the Parameter Report (1998) and most of those parameterswere identified to DOE in the aforementioned Agency letter of 19 March 1997 (Trovato 1997A).However, additional parameters or groups of parameters were added to the analysis based on theinitial results of the Agency’s sensitivity studies and on concerns for specific parameters andprocesses expressed during EPA’s public hearings and in public written comments.

This third report, Technical Support Document for Section 194.23 - Parameter JustificationReport [Docket A-93-02, Item V-B-14], is referred to as the Justification Report (JR). Itdescribes the disposition of the inadequately supported parameters described in the Agency’sletter of March 19, 1997 (Trovato 1997A). This disposition was based on the results of theAgency’s sensitivity analysis, additional supporting information provided by the DOE, andfurther analysis by the Agency. Parameters were removed from the list by the Agency if, forexample, PA performance measures were found to be insensitive to them, if the additional DOEsupporting information was found to be adequate, or if upon further review the Agency

2

determined that DOE’s existing supporting rationale was acceptable. The disposition of theseparameters was described to DOE in the Agency’s letters of 17 April 1997 (Trovato 1997B) and25 April 1997 (Trovato 1997C).

Parameters that were not removed from the list were used in developing a revised data base ofparameters of major concern to the Agency for use in the EPA-mandated PerformanceAssessment Verification Test (PAVT). The rationale supporting development of this revised database is presented in this report. The PAVT is designed to provide a comprehensive test of theeffects of changes in significant, uncertain parameters and changes in other aspects of the CCAPA computer codes on the PA compliance calculations presented by DOE in the CCA.

1.2 Report Structure

This report is divided into six sections. Following this introduction, an overview of the parameterselection process is presented in Section 2. This overview describes the Agency’s review ofparameter documentation, the screening process, and the sensitivity analysis process that wereused to disposition the inadequately supported parameters. Sections 3, 4, and 5 present detaileddiscussions of the inadequately supported parameters listed in Enclosures 2, 3, and 4,respectively, of the Agency’s aforementioned letter of March 19, 1997 (Trovato 1997A). Finally,Section 6 presents a summary of the results of the Agency’s analysis of those parameters.

3

2.0 PARAMETER SELECTION PROCESS

2.1 Parameter Selection

The parameters identified to DOE as being inadequately supported were the result of a screeningprocess described in the Parameter Report (1998). In overview, this process began with adetailed EPA review of the Sandia National Laboratories (SNL) parameter data base inAlbuquerque, New Mexico. Individual Records Packages were obtained from SNL’s NuclearWaste Management Program Information Service Center and reviewed to identify the source andrationale for the principal parameters used in the CCA PA model. Deficiencies in documentationand lack of adequate rationale were identified and used along with information from the CCAand past SNL/DOE reports to begin an evaluation of parameter importance. This effortproceeded as a process of iterative sifting of the original CCA PA parameter database to identifythose parameters deemed important to the performance of the WIPP.

The initial review involved nearly 1600 parameters. Of these, about 465 parameters were foundto be worthy of a more detailed evaluation. These 465 parameters were given additionalscreening for importance and uncertainty, and about 150 were considered to have potentialimpact on performance. These 150 parameters were further reviewed and about 60 parameterswere identified for inclusion in EPA’s March 19, 1997 letter (Trovato 1997A). Those parameterswere grouped by type of deficiency and are listed in Table 2.1-1.

2.2 Parameter Review

In the Agency’s letter of March 19, 1997 (Trovato 1997A) the DOE was told that the issuesdescribed included EPA’s outstanding concerns with the CCA and was requested to resolvethose concerns. DOE responded by providing additional documentation and rationale which theAgency found helpful in evaluating many of the parameters. Additional information on many ofthe parameters was developed through technical reviews by the Agency and was also used tohelp in the evaluation process. Details of this additional documentation and rationale and thedecisions they supported are presented in the parameter-by-parameter discussions in Sections 3through 5 of this report.

Additionally, many of the inadequately supported parameters identified in the Agency’s March19, 1997 letter (Trovato 1997A) were selected by the Agency for detailed sensitivity analysis.These parameters, supplemented by others that were added later, totaled about 80 inputs thatwere evaluated in that analysis . PA model sensitivity to changes in both individual inputparameters and groups of parameters was evaluated in 40 sensitivity analyses performed by theAgency. In most of these analyses, the parameters for which sensitivity was being analyzed wereassigned low, high, and baseline values. The low and high values were selected by EPA toprovide an indication of the changes in model output performance measures resulting from largeand sometimes extreme changes, both up and down, in the input parameter values. Theperformance measures are model output parameters selected by the Agency that either directlyrepresent or strongly influence radionuclide releases calculated in performance assessment and

4

therefore serve as indicators of the importance of uncertainty in an input parameter to repositoryperformance. An example performance measure is the calculated direct brine release, whichdirectly represents a radionuclide release and is an output parameter of the BRAGFLO-DBRmodel. Another example performance measure is the calculated repository gas pressure, which isan output parameter of the BRAFGLO model and strongly influences spallings and brinereleases during inadvertent borehole intrusions. In most cases, a percent change in a performancemeasure will result in a much smaller percent change in the total radionuclide release. Theselection of performance measures for the SA is more fully discussed in Appendix PM of the SAReport (1998).

The baseline values for the sensitivity analysis were generally the values used by DOE in theCCA PA model or, in the case of sampled parameters, the median of the sampled range. Thesensitivity analysis results were analyzed by calculating the average absolute percent change inthe performance measure for the high and low parameter values. The results were found to rangefrom zero to many thousand percent and were used to help select the inadequately supportedparameters. The detailed results of this analysis are presented in the SA Report (1998).

Using additional information of the type described above, the Agency was able to conclude thatmany of the parameters were no longer in question. The supporting reasons for this findingincluded evidence from the SA that the PA model was not sensitive to changes in the value ofthe parameter, presentation by DOE of adequate supporting data for the parameter value, ordevelopment of adequate justification for the parameter value or range by DOE or through theAgency’s own studies. Those parameters that were not included in this category were givenalternative values that were considered by the Agency to be more appropriate. The Agencyrequired DOE to use these alternative values in the PAVT, as stated in the Agency’s letters of 17April 1997 (Trovato 1997B) and 25 April 1997 (Trovato 1997C). For those parameters that werechanged in the PAVT, the Agency’s rationale for selecting the mandated alternative values ispresented in Sections 3 through 5 of this report.

5

Table 2.1-1Inadequately Supported Parameters Identified in EPA’s March 19, 1997 Letter

No. Parameter No. Material ID - Parameter ID Parameter Description.A. Parameters Lacking Supporting Data (Enclosure 2 Parameters)

1 3245 BLOWOUT - CEMENT Waste Cementation Strength2 3246 BLOWOUT - PARTDIA Waste Particle Diameter3 198 DRZ_1 - PRMX_LOG DRZ Permeability4 2177 S_MB_139 - DPHIMAX Incremental Increase in Anhydrite Porosity in MB

1395 2180 S_MB_139 - PF_DELTA Incremental Pressure for Full Fracture Development6 586 S_MB_139 - PI_DELTA Fracture Initiation Pressure Increment7 2178 S_MB_139 - KMAXLOG Maximum Permeability in Altered Anhydrite8 3134 BH_OPEN - PRMX_LOG Open Borehole Permeability9 2158 S_ANH_AB - DPHIMAX Incremental Increase in Anhydrite Porosity in Beds A

and B10 214 EXP_AREA - PRMX_LOG Experimental Area Permeability11 3473 BLOWOUT - THCK_CAS Thickness of Castile Brine Pocket for Direct Brine

Release12 3456 BLOWOUT - RE_CAST Radius of Castile Brine Pocket for Direct Brine

Release13 2918 CASTILER - GRIDFLO Index for Selecting Brine Pocket Volume

B. Parameters with Different Values or Ranges (Enclosure 3 Parameters)

1 3493 GLOBAL - PBRINE Probability of Encountering Pressurized Brine2 2254 BOREHOLE - TAUFAIL Waste Shear Resistance3 3184 BH_SAND - PRMX_LOG Long-Term Borehole Permeability4 2918 CASTILER - VOLUME Castile Brine Pocket Volume5 61 CASTILER - COMP_RCK Castile Brine Pocket Rock Compressibility

C. Parameters with Questionable Values or Ranges (Enclosure 4 Parameters)

1 27 BOREHOLE - DOMEGA Drill String Angular Velocity2 64 CASTILER - POROSITY Castile Brine Pocket Porosity3 66 CASTILER - PRESSURE Castile Brine Pocket Pore Pressure4 259 PAN_SEAL - PRMX_LOG Panel Seal Permeability5 528 S_ANH_AB - POROSITY Effective Porosity of Anhydrite Beds A and B6 567 S_MB138 - POROSITY Effective Porosity of Anhydrite MB 1387 588 S_MB139 - POROSITY Effective Porosity of Anhydrite MB 1398 651 WAS_AREA - ABSROUGH Waste Area Absolute Roughness9 653 WAS_AREA - COMP_RCK Waste Area Rock Compressibility10 1992 WAS_AREA - DIRNCCHW Bulk Density of Iron Containers in CH Waste11 1993 WAS_AREA - DIRNCRHW Bulk Density of Iron Containers in RH Waste12 2040 WAS_AREA - DIRNCHW Average Density of Iron-Based Material in CH Waste13 2041 WAS_AREA - DCELLCHW Average Density of Cellulosics in CH Waste

Table 2.1-1Inadequately Supported Parameters Identified in EPA’s March 19, 1997 Letter

(Continued)

6

No. Parameter No. Material ID - Parameter ID Parameter Description

14 2274 WAS_AREA - DCELLRHW Average Density of Cellulosics in RH Waste15 2907 STEEL - CORRMCO2 Steel Corrosion Rate16 3147 CONC_PLG - POROSITY Borehole Plug Porosity17 3185 CONC_PLG - PRMX_LOG Borehole Plug Permeability18 3256 BLOWOUT - FGE Gravity Scaling Factor19 3259 BLOWOUT - APORO Waste Permeability in CUTTINGS_S Model20 3429 PHUMOX3 - PHUMCIM Humic Colloid Proportionality Constant in Castile

Brine21 3471 BLOWOUT - MAXFLOW Maximum Period of Uncontrolled Borehole Flow22 3472 BLOWOUT - MINFLOW Minimum Period of Uncontrolled Borehole

Flow23 3433 PHUMOX3 - PHUMSIM Humic Colloid Proportionality Constant in Salado

Brine24 3470 BLOWOUT - GAS_MIN DBR Cutoff Gas Flow Rate25 3317 PU - PROPMIC Microbial Colloid Proportionality Constant for

Plutonium26 3405 SOLMOD6 - SOLCIM U(VI) Solubility Limit in Castile Brine27 3409 SOLMOD6 - SOLSIM U(VI) Solubility Limit in Salado Brine28 3402 SOLMOD3 - SOLCIM Oxidation State +3 Solubility Limit in Castile Brine29 3403 SOLMOD4 - SOLCIM Oxidation State +4 Solubility Limit in Castile Brine30 3407 SOLMOD4 - SOLSIM Oxidation State +4 Solubility Limit in Salado Brine31 3404 SOLMOD5 - SOLCIM Oxidation State +5 Solubility Limit in Castile Brine32 34-8 SOLMOD5 - SOLSIM Oxidation State +5 Solubility Limit in Salado Brine33 3311 AM - PROPMIC Microbial Colloid Proportionality Constant for

Americium34 3482 AM+3 - MKD_AM Matrix Partition Coefficient for Americium +335 3480 PU+3 - MKD_PU Matrix Partition Coefficient for Plutonium +336 3481 PU+4 - MKD_PU Matrix Partition Coefficient for Plutonium +437 3479 U+4 - MKD_U Matrix Partition Coefficient for Uranium +438 3475 U+6 - MKD_U Matrix Partition Coefficient for Uranium +639 656 WAS_AREA - GRATMICH Gas Generation Rate due to Microbial Action under

Humid Conditions40 657 WAS_AREA - GRATMICI Gas Generation Rate due to Microbial Action under

Inundated Conditions

7

3.0 PARAMETERS LACKING SUPPORTING DATA

Parameters for which the Agency was initially unable to find supporting data are addressed inthis section. These parameters are listed in Enclosure 2 of the Agency’s March 19, 1997 letter(Trovato 1997A). As stated in that letter, the Agency considers the traceability of the parameterand data record packages that support the input parameter values used in the PA to be important,and requires that the critical input parameters must either be supported by actual data collectionand/or results of experimentation, or be based on expert judgement following the Agency’sprocedures. Each of the following subsection headings identifies the parameter’s referencenumber, Material ID, and Parameter ID, followed by a brief description of the parameter.Parameter values are often presented in the format used for model input and very large or verysmall numbers may be expressed as logarithms. For example, the permeability 1.7 x 10-13 m2 maybe expressed as log10 1.7 E-13 m2 or -12.769 log m2.

3.1 3245 BLOWOUT - CEMENT: Waste Cementation Strength

The waste cementation strength was used in the CUTTINGS_S code for predicting spallingsreleases. It is a measure of the strength imparted to the waste by cementitious materials that areexpected by DOE to form in waste pores due to the precipitation of salts in the repository brine.The cement is expected to be primarily derived from precipitation and redeposition of halite,MgO reaction products, and corrosion reaction products around waste particles during the life ofthe repository. This parameter is not supported by data collection or experimentation, nor was itdeveloped by expert judgement following the Agency’s procedures (see WPO # 38241). Thisparameter was reviewed by DOE’s Engineered Systems Peer Review Panel, which stated that itsvalue was highly uncertain and could range from near zero to 700 psi and recommended treatingit as a sampled variable in the PA (Ross-Brown et al. 1996 p. 16). Additionally, at the time theAgency’s March and April letters to DOE were prepared (Trovato 1997A; 1997B; 1997C), thespallings release model had not been approved by the DOE’s Conceptual Models Peer ReviewPanel (Wilson et al. 1997A p. 2ff). Because of its uncertainty and potential significance incomputing spallings releases, this parameter was identified for further evaluation.

Waste cementation strength was included in the Agency’s sensitivity analysis and given a highvalue that was 700 times the value used in the CCA PA. The average absolute change for thisparameter was found to be 149%. Although this sensitivity is low compared with results forother parameters, it exceeded the Agency’s threshold value of 25%. Partially because of thedemonstrated sensitivity of spallings releases to this parameter, but primarily because of theimportance of spallings releases relative to total repository releases, a lack of confidence in thespallings release calculations, a lack of documented support for the adopted parameter value, andthe high degree of uncertainty surrounding the correct parameter value, waste cementationstrength was retained as a mandated, sampled variable for the PAVT (Trovato 1997C Enclosure2).

For the PAVT, waste cementation strength was mandated in the Agency’s letter of April 25,1997 to be treated as a sampled variable with a log-uniform distribution, a minimum value equal

8

to the minimum value to be set for the waste shear resistance (parameter BOREHOLE -TAUFAIL), and a maximum value of 4.8 E+06 Pa (see Trovato 1997C Enclosure 2). A log-uniform distribution was selected for this parameter because, based on present knowledge, theAgency believes that all values across the range of several orders of magnitude are equallylikely. The minimum value was set equal to the minimum value of the waste shear resistanceused in CUTTINGS_S to calculate cavings releases, based on the rationale that both parametersare a measure of the resistance of waste to the shear forces imposed by a fluid moving over thewaste surface and there was no evident reason why the minimum value of the two parametersshould be different. The maximum value was set equal to the upper end of the strength rangerecommended by the aforementioned Engineered Systems Peer Review Panel and is equal to 700psi, which is the approximate strength of sorel cement.

DOE’s Conceptual Models Peer Review Panel issued its final report at about the time that theforegoing data requirements were transmitted by the Agency to DOE. In that report the Panelcontinued to find the spallings conceptual model to be inadequate, but determined that thespallings volumes used by DOE in the CCA (0.5 to 4.0 m3 per spall event) were reasonable forpurposes of performance assessment (Wilson et al. 1997B p. 12). In view of this, DOE proposedand the Agency accepted that for the PAVT the spallings model would be changed to directlysample releases across the Panel’s approved range from a uniform distribution rather thancalculate those releases with the computational model. Although this change eliminated the needto vary the waste cementation strength parameter in the PAVT, its acceptance by the Agencywas conditioned on the demonstration that the waste particle diameter distribution used by DOEin the CCA is more conservative (that is, addresses smaller particles) than the distributionresulting from expert judgement (see Section 3.2 below).

3.2 3246 BLOWOUT - PARTDIA: Waste Particle Diameter

The waste particle diameter was used in the CUTTINGS_S model to calculate spallings releasesand to determine the waste shear resistance for calculating cavings releases. The waste particlediameter was identified in the Agency’s letter of March 19, 1997 as lacking supporting evidence(Trovato 1997A Enclosure 2), and was further discussed in the Agency’s letter of April 25, 1997as not being supported by data and therefore requiring derivation through “expert judgement”(Trovato 1997C p. 1). The Agency considers expert elicitation (or expert judgement) to be anappropriate method for obtaining values of parameters that are not readily measurable or arehighly uncertain and are therefore not supported by an extensive database. In summary, themethod consists of convening a panel of acknowledged experts in a variety of fields associatedwith aspects of the parameters of concern who each contribute toward a final judgementregarding the parameter values. The advantage of the method is that for such parameters, thevalues can be selected to reflect probable sources of error in measurement, modeling, and the useof analogs from other sites, and the ranges can be selected to reflect the uncertainties in theparameter values. The alternative of basing parameter values only on available data without thebenefit of judgement requires an extensive data base to assure that the values and associateduncertainties are adequately represented. In response to the Agency’s mandate, the DOEconvened an Expert Elicitation Panel (EEP) to derive values for waste particle diameters

9

appropriate for use in WIPP PA. The results of that elicitation process are documented in theEEP’s final report (DOE Carlsbad Area Office 1997).

DOE’s analysis of the EEP’s findings are summarized in an August 5, 1997 memorandum (WPO# 46936), wherein a bounding calculation is described that correlates the distributions derived bythe EEP to the mean particle diameters used in the cavings and spallings models. Converting theEEP’s particle size distributions to a volume fraction basis and considering complete wastedegradation as a worst case, DOE calculated a lower bound mean diameter of 0.1 cm and anupper bound mean diameter ranging from 10 cm (the initial mean particle size, assuming nocementation) to the size of the waste panel (assuming that reacted MgO precipitates as acement). The particle diameters were sampled from a log-uniform distribution in the CCA PAand ranged from 0.004 cm to 20 cm with a median value of 0.28 cm. The Agency acceptedDOE’s analysis as adequately demonstrating that the waste particle diameter distribution used inthe CCA addresses smaller particles and is therefore more conservative than the distributionresulting from expert judgement. Calculations based on smaller particle sizes are moreconservative because they result in greater radionuclide releases. Based on this determination,the Agency accepted the spallings model change described in Section 3.1, which eliminated theuse of particle size information in determining spallings releases.

As previously mentioned, the waste particle diameter is used in the cuttings release calculationsto determine waste shear resistance (parameter BOREHOLE - TAUFAIL). That parameter wasidentified as being inadequately supported in Enclosure 3 of the Agency’s letter of March 19,1997 (Trovato 1997A). Use of the EEP’s waste particle size distributions in developing alternatevalues for that parameter is discussed in Section 4.2 of this report.

3.3 198 DRZ_1 - PRMX_LOG: DRZ Permeability

The DRZ permeability is the permeability in the X-direction of the disturbed rock zone (DRZ)surrounding the repository. This parameter, entered into the PA database as a logarithm, wasassumed to be a constant equal to -15 log m2 in the CCA PA. In Chapter 6, Section 6.4.5.3 DOEstates that using a constant DRZ permeability is conservative. This parameter is potentiallysignificant in computing spallings and direct brine releases because of its influence on repositorygas pressure buildup. It was identified in the Agency’s March 19, 1997 letter for furtherevaluation (Trovato 1997A Enclosure 2).

The DRZ permeability was included in the Agency’s sensitivity analysis where it was given ahigh value of 12.5 log m2 and a low value of -21 log m2 (see SA Report 1998 Appendix PDSection PD-3.12). These values were determined by the Agency based on the maximum andminimum of 14 field permeability measurements within the DRZ of the WIPP excavations (seeWPO # 32038 Rev 1 p. 2). The average absolute change in the performance measures for thisparameter was found to be 325%, which the Agency considered potentially significant (see SAReport 1998 Table 3.1-1). Because of the demonstrated sensitivity of the PA model results tothis parameter and because of the potential importance of this parameter in calculating gaspressure buildup in the repository, alternate values were identified for this parameter and its Y-

10

and Z-direction counterparts in the Agency’s April 17, 1997 letter for use in the PAVT (Trovato1997B Enclosure 2).

The DRZ log permeabilities specified by the Agency for use in the PAVT ranged from a low of -19.4 log m2 to a high of -12.5 log m2, to be sampled from a uniform distribution with a medianof -15.95 log m2 (see Trovato 1997B Enclosure 2 and Kruger 1997). The low end of the range isbased on the lowest value measured by DOE in gas permeability tests in anhydrite cores fromMarker Bed 139 (see WPO # 30603 p. 2 and WPO # 32038). The test samples were disturbed bythe drilling and sample preparation processes (see WPO # 38367 p. 11) and the Agency believesthat the lowest of those measured permeabilities provides an appropriate lower bound forsampling DRZ permeability in the PAVT. The high end of the range is the same as the highvalue of the sensitivity analysis and is based on the reasoning described above. A uniformdistribution was selected for the log-permeabilities because, based on present knowledge, theAgency believes that all values across the range of several orders of magnitude are equallylikely. The median value is determined from the type and range of the distribution. Thismandated approach is considered by the Agency to adequately reflect both the uncertainty andthe reasonable range of values for this parameter. The variation of this parameter in the PAVT isexpected to result in a greater range of computed spallings and direct brine releases because ofits influence on repository gas pressure buildup. The lower median permeability in the PAVT isexpected to conservatively result in higher direct brine releases because the restricted gasmigration may result in higher repository gas pressures.

3.4 2177 S_MB_139 - DPHIMAX: Maximum Incremental Increase in AnhydritePorosity in Marker Bed 139

Porosity was allowed to increase in the anhydrite interbeds near the repository in response toincreases in gas/brine pressure in the repository. Parameter S_MB_139 - DPHIMAX was used toestablish the maximum incremental porosity increase that would be allowed in anhydrite MarkerBed 139. It was treated as a constant in the CCA PA equal to 0.039. In its initial review, theAgency did not find this parameter to be supported by data collection or experimentationdocumented in the CCA or SNL WIPP records center, nor was it developed by expert judgementfollowing the Agency’s procedures. This parameter is potentially significant in computing brinereleases through MB 139 to the accessible environment and was identified in the Agency’sMarch 19, 1997 letter for further evaluation (Trovato 1997A Enclosure 2).

Anhydrite porosity was selected by the Agency for sensitivity analysis to determine whether PAmodel results were sensitive to changes in this and other parameters used in determining thevalue of that porosity. In that sensitivity analysis, DOE’s initial, undisturbed anhydrite porosityof 0.011 was given a low value of 0.006 and a high value of 0.017 in all interbeds. These valueswere determined by the Agency based on the minimum and maximum porosities measured byDOE in 16 samples collected from the WIPP Site (see WPO # 34860 p. 2). The results of thesensitivity analysis showed no sensitivity of the model performance measures to changes in thisparameter (see SA Report 1997 Table 3.1-1). Because of the lack of sensitivity to changes inanhydrite porosity for the range of values defined by the Agency in this analysis, additional

11

sensitivity studies addressing the higher maximum porosity of 0.05 (equal to 0.011 + 0.039)defined by the parameter S_MB_139 - DPHIMAX were not considered necessary.

In addition to the results of the sensitivity analysis for anhydrite porosity, in a letter dated April15, 1997 (Docket: A-93-02, II-I-24, Comment No. 9, Larson et al.1997) the Agency receivedadditional information from DOE describing the use of experimental data in developing theanhydrite fracture generation parameters including those discussed in Sections 3.4 through 3.7and 3.9 of this report. This letter and its attachments provided a list of the relevantcontemporaneous documentation and a summary that provided a historical perspective of thedevelopment of the conceptual model and its implementing parameters and data base. Uponreview, this information was found to provide an acceptable explanation of the experimentalbasis for the anhydrite fracture generation parameters addressed in this report.

As part of its review of this and the related anhydrite interbed parameters discussed in Sections3.5 through 3.7 and 3.9 below, the Agency evaluated the appropriateness of the pressure-dependent porosity model used in the CCA compared with an alternative pressure-dependentaperture model (see Freeze et al. 1995 p. 6-25). The porosity model treats the anhydrite interbedas an equivalent porous medium and simulates the effect of fracturing by increasing bothporosity (to increase storage) and permeability (to increase flow). DOE structured the porositymodel with exponential terms to simulate the rapid rise in permeability accompanying fracturing.The aperture model addresses both increasing storage and flow by directly simulatingpropagation of an open fracture. The two models are capable of producing similar solutions forpermeability changes and both have been used to match field data on permeability versus porepressure (see Beauheim et al.1994 Figure 20). However, the models do not provide similarpermeability/porosity correlations and for a given porosity change the aperture model predicts ahigher permeability than the porosity model. Based on the pervasive presence of natural fracturesin the anhydrite interbeds and on the results of hydrofracturing tests conducted at the WIPP Siteby DOE, the Agency believes that the porosity model better represents conditions in the WIPPinterbeds than the aperture model (see Beauheim et al. 1993). In view of DOE’s use of field andexperimental data to select the model parameters, the Agency considers the application of theporosity model in the CCA to be reasonable.

In view of the acceptability of the additional information provided by DOE and the lack ofmodel sensitivity to changes in anhydrite porosity, this parameter was identified in the Agency’sApril 25, 1997 letter as being no longer in question (Trovato 1997C Enclosure 1).

12

3.5 2180 S_MB_139 - PF_DELTA: Incremental Pressure for Full FractureDevelopment in Marker Bed 139

Permeability and porosity were both allowed to increase in the anhydrite interbeds near therepository in response to increases in gas/brine pressure in the repository. Although fracturedevelopment was not specifically modeled, the process was conceptually considered to beanalogous to hydrofracturing. This parameter was used to establish the incremental increase inrepository pressure in Marker Bed 139 above the initial pressure in the anhydrite that wouldcorrelate to the maximum allowable permeability and porosity values. This parameter wastreated as a constant in the CCA PA equal to 3.8 E+06 Pa. In its initial review, the Agency didnot find this parameter to be supported by data collection or experimentation, nor was itdeveloped by expert judgement following the Agency’s procedures. This parameter is potentiallysignificant in computing brine releases through MB 139 to the accessible environment and wasidentified in the Agency’s March 19, 1997 letter for further evaluation (Trovato 1997AEnclosure 2).

To assess the sensitivity of the PA model to changes in this parameter, the Agency performedthree sensitivity analyses that independently changed anhydrite porosity, permeability, andpressure. The sensitivity analysis for anhydrite porosity is discussed in Section 3.4 above andindicated no sensitivity of the model performance measures to changes in porosity.

The sensitivity analysis for anhydrite permeability was performed using a low value of -12 logm2 and a high value of -17.1 log m2 (see SA Report 1998 Appendix PD Section PD-1.1).Anhydrite permeability was treated by DOE as a sampled variable in the CCA with a Student Tdistribution, a median of -18.89 log m2, and a range equal to the high and low values used by theAgency in its sensitivity analysis. These values were based on direct field measurements (seeCCA, Volume XI, Appendix PAR p. PAR-81 [Docket: A-93-02, II-G-1]). This analysis showeda 2347% average change in the performance measures, which the Agency considers to besignificant (see SA Report 1998 Table 3.1-1).

The sensitivity analysis for anhydrite initial pressure was performed using a low value of 1.1E+07 Pa and a high value of 1.38 E+07 Pa (see SA Report 1998 Appendix PD Section PD-1.3).Anhydrite pressure was treated by DOE as a sampled variable in the CCA with a uniformdistribution and a range equal to the high and low values used by the Agency in its sensitivityanalysis. These values were based on direct field measurements (see CCA, Volume XI,Appendix PAR p. PAR-99 [Docket: A-93-02, II-G-1]). This analysis showed a 10% averagechange in the performance measures, which the Agency does not consider to be significant (seeSA Report 1998 Table 3.1-1). For most parameters evaluated in the sensitivity analysis includingthe anhydrite initial pressure, the percent change in the overall radionuclide release will beconsiderably less than the percent change in the performance measures. The performancemeasures used in the sensitivity analysis are fully described in the SA Report.

Of the three parameters related to parameter S_MB_139 - PF_DELTA that were studied insensitivity analysis, the model performance measures were only sensitive to changes in anhydrite

13

permeability. Despite this demonstrated sensitivity, anhydrite permeability was not identified bythe Agency as requiring changed values in the PAVT because the anhydrite permeability valuesupon which DOE based its analysis are extensive, appropriate, well documented, and werecollected under an NQA-1 QA program. They are based on 5 in situ hydraulic tests and 31laboratory tests from anhydrite interbeds in the Salado, and screened for both scale effects andrepository-induced disturbance (see CCA, II-G-1, Volume XI, Appendix PAR p. PAR-81[Docket: A-93-02, II-G-1]). The Agency considers DOE’s treatment of anhydrite permeability inthe CCA to adequately capture the expected value and uncertainty in this parameter.

In addition to the results of the foregoing sensitivity analyses, in a letter dated April 15, 1997 theAgency received additional information from DOE ( Larson et al.1997) describing the use ofexperimental data in developing the anhydrite fracture generation parameters including thosediscussed in Sections 3.4 through 3.7 and 3.9 of this report. This letter is further discussed inSection 3.4 and was found to provide an acceptable explanation of the experimental basis for theanhydrite fracture generation parameters addressed in this report.

In view of the acceptability of the additional information provided by DOE, the results of thesensitivity analyses, and the acceptability of DOE’s anhydrite permeability database, theparameter S_MB_139 - PF_DELTA was identified in the Agency’s April 25, 1997 letter asbeing no longer in question (Trovato 1997C Enclosure 1).

3.6 586 S_MB_139 - PI_DELTA: Fracture Initiation Pressure Increment in MarkerBed 139

As described in Section 3.5, permeability and porosity were both allowed to increase in theanhydrite interbeds near the repository in response to increases in gas/brine pressure in therepository. This parameter was used to establish the minimum incremental increase in repositorypressure above the initial pressure in the anhydrite that would initiate increases in thepermeability and porosity of Marker Bed 139. This parameter was treated as a constant in theCCA PA equal to 0.2 E+06 Pa. In its initial review, the Agency did not find this parameter to besupported by data collection or experimentation, nor was it developed by expert judgementfollowing the Agency’s procedures. This parameter is potentially significant in computing brinereleases through MB 139 to the accessible environment and was identified in the Agency’sMarch 19, 1997 letter for further evaluation (Trovato 1997A Enclosure 2).

The Agency evaluated the sensitivity of PA model performance measures to changes inanhydrite interbed pressure by assigning low and high values to the initial value of that pressure.This sensitivity analysis is described in Section 3.5 above and indicated only a minor sensitivityto changes in this parameter. In addition to the results of the foregoing sensitivity analysis, in aletter dated April 15, 1997 the Agency received additional information from DOE (Larson etal.1997) describing the use of experimental data in developing the anhydrite fracture generationparameters including those discussed in Sections 3.4 through 3.7 and 3.9 of this report. Thisletter is further discussed in Section 3.4 and was found to provide an acceptable explanation ofthe experimental basis for the anhydrite fracture generation parameters addressed in this report.

14

In view of the acceptability of the additional information provided by DOE and the results of thesensitivity analysis, the parameter S_MB_139 - PI_DELTA was identified in the Agency’s April25, 1997 letter as being no longer in question (Trovato 1997C Enclosure 1).

3.7 2178 S_MB_139 - KMAXLOG: Maximum Permeability in Altered Anhydrite inMarker Bed 139

As described in Section 3.5, permeability was allowed to increase in the anhydrite interbeds nearthe repository in response to increases in gas and brine pressure in the repository. This parameterwas used to establish the maximum log permeability in Marker Bed 139 that the model wouldallow to occur in response to increased repository pressures. This parameter was treated as aconstant in the CCA PA equal to -9 log m2. In its initial review, the Agency did not find thisparameter to be supported by data collection or experimentation, nor was it developed by expertjudgement following the Agency’s procedures. This parameter is potentially significant incomputing brine releases through MB 139 to the accessible environment and was identified inthe Agency’s March 19, 1997 letter for further evaluation (Trovato 1997A Enclosure 2).

In a letter dated April 15, 1997 the Agency received additional information from DOE ( Larsonet al.1997) describing the use of experimental data in developing the anhydrite fracturegeneration parameters including those discussed in Sections 3.4 through 3.7 and 3.9 of thisreport. This letter is further discussed in Section 3.4 and was found to provide an acceptableexplanation of the experimental basis for the anhydrite fracture generation parameters addressedin this report.

In view of the acceptability of the additional information provided by DOE this parameter wasidentified in the Agency’s April 25, 1997 letter as being no longer in question (Trovato 1997CEnclosure 1).

3.8 3134 BH_OPEN - PRMX_LOG: Open Borehole Permeability

The open borehole permeability is the permeability assigned to the unplugged parts of anintrusion borehole during the 200-year period before the borehole is assumed to be filled withcasing corrosion products and material sloughed from the borehole walls. This parameter wastreated as a constant in the CCA PA equal to -9 log m2. In its initial review, the Agency did notfind this parameter to be supported by data collection or experimentation, nor was it developedby expert judgement following the Agency’s procedures. This parameter is potentiallysignificant in computing long term brine releases through an intrusion borehole to the groundsurface and was identified in the Agency’s March 19, 1997 letter for further evaluation (Trovato1997A Enclosure 2).

The open borehole permeability was one of the parameters selected by the Agency for sensitivityanalysis. In that analysis, DOE’s log permeability value was increased to -6 log m2 . TheAgency’s objective was to determine the effect of a higher permeability value, and the low valuewas therefore set equal to the baseline CCA value. The SA results showed no sensitivity to this

15

three order of magnitude increase in open borehole permeability (see SA Report 1998 Table 3.1-1). Because of a lack of sensitivity, this parameter was identified in the Agency’s April 25, 1997letter as being no longer in question (Trovato 1997C Enclosure 1).

3.9 2158 S_ANH_AB - DPHIMAX: Incremental Increase in Anhydrite Porosity in BedsA and B

This parameter applies to anhydrite beds A and B and is the same as defined for Marker Bed 139in Section 3.4. It was treated as a constant in the CCA PA equal to 0.239. This value wasselected by DOE to maintain equal transmissivities in all anhydrite interbeds in the porousmedium model during the conceptual hydrofracturing process. Through the porosity-permeability correlation in the model it accounts for differences in thickness between anhydritebeds A and B and Marker Bed 139. The value of this parameter is therefore tied to the value ofthe parallel parameter for Marker Bed 139. In its initial review, the Agency did not find thisparameter to be supported by data collection or experimentation, nor was it developed by expertjudgement following the Agency’s procedures. This parameter is potentially significant incomputing brine releases through anhydrite beds A and B to the accessible environment and wasidentified in the Agency’s March 19, 1997 letter for further evaluation (Trovato 1997AEnclosure 2). Because of its relationship to the aforementioned parallel parameter for MarkerBed 139, the Agency found parameter S_ANH_AB - DPHIMAX to be no longer in question forthe reasons given for parameter S_MB_139 - DPHIMAX in Section 3.4. This finding isdocumented in the Agency’s April 25, 1997 letter (Trovato 1997C Enclosure 1).

3.10 214 EXP_AREA - PRMX_LOG: Experimental Area Permeability

The WIPP repository excavations may be divided into the waste disposal area, the operationsarea, and the experimental area. The experimental area permeability is the permeability assignedin the PA model to the experimental area of the repository. This area and the operations area willnot be filled with waste and will be allowed to close naturally through halite creep uponrepository closure. This parameter was treated as a constant in the CCA PA equal to -11 log m2

and was the same as the permeability assigned to the operations area. In its initial review, theAgency did not find this parameter to be supported by data collection or experimentation, norwas it developed by expert judgement following the Agency’s procedures.

The experimental and operations area permeabilities were varied together in a sensitivityanalysis conducted by the Agency. In that analysis, DOE’s log permeability value was decreasedto -17 log m2. The Agency’s objective was to determine the effect of a lower permeability value,and the high value was therefore set equal to the baseline CCA value. The SA results showed nosensitivity to this six order of magnitude decrease (see SA Report 1998 Table 3.1-1). Because ofa lack of sensitivity, this parameter was identified in the Agency’s April 25, 1997 letter as beingno longer in question (Trovato 1997C Enclosure 1).

3.11 3473 BLOWOUT - THCK_CAS: Thickness of Castile Brine Pocket for Direct BrineRelease

16

The parameters BLOWOUT - THCK_CAS and BLOWOUT - RE_CAST (see Section 3.12)were used in the PA model to define the thickness and radius, respectively, of a Castile brinepocket for the Direct Brine Release (DBR) model. The Castile brine pocket volume was treatedas a constant in the CCA PA with an equivalent thickness equal to 12.34 m and an equivalentradius equal to 114 m. The thickness is based on the estimated thickness of the Castile anhydritebeds underlying the site that are believed by DOE to potentially contain brine pockets, and theequivalent radius of 114 m approximates the area of a waste panel. In its initial review, theAgency did not find this parameter to be supported by data collection or experimentation, norwas it developed by expert judgement following the Agency’s procedures. During this reviewthe Agency found that the equivalent radius used in the CCA was in error and that 230 mprovided a more appropriate estimate of the actual area of a waste panel as documented by DOEin the documentation on direct brine release calculations. This parameter is potentiallysignificant in computing direct brine releases through intrusion boreholes to the ground surfaceand was identified in the Agency’s March 19, 1997 letter for further evaluation (Trovato 1997AEnclosure 2).

The parameters BLOWOUT - THCK_CAS and BLOWOUT - RE_CAST were varied togetherin a sensitivity analysis conducted by the Agency. In that analysis, the low brine pocket volumewas assigned a thickness of 24 m and an equivalent radius of 30 m. These correspond to a grossvolume of about 68,000 m3. A thickness of 24 m is the maximum thickness of the Castileanhydrite beds believed by DOE to potentially contain brine pockets, and an equivalent radius of30 m approximates the area of a single waste disposal room and surrounding salt pillars. Thehigh brine pocket volume was assigned a thickness of 7 m and an equivalent radius of 230 m,corresponding to a gross volume of about 1,163,000 m3. A thickness of 7 m is the minimumthickness of the Castile anhydrite beds believed by DOE to potentially contain brine pockets, andbased on the corrected equivalent radius of 230 m approximates the area of two waste panels.For comparison, DOE’s CCA PA dimensions correspond to a gross volume of about 500,000 m3.By coupling the largest thickness with the smallest radius for the low volume, and the smallestthickness with the largest radius for the high volume, a range of volumes was achieved that theAgency considers appropriate for evaluating the sensitivity of DBR model results to changes inbrine pocket volume. The SA results showed no sensitivity to these volume changes (see SAReport 1998 Table 3.1-1). Because of a lack of sensitivity, both BLOWOUT - THCK_CAS andBLOWOUT - RE_CAST were identified in the Agency’s April 25, 1997 letter as being nolonger in question (Trovato 1997C Enclosure 1).

3.12 3456 BLOWOUT - RE_CAST: Radius of Castile Brine Pocket for Direct BrineRelease

This parameter is discussed along with BLOWOUT - THCK_CAS in Section 3.11. Because of alack of sensitivity, both BLOWOUT - THCK_CAS and BLOWOUT - RE_CAST were identifiedin the Agency’s April 25, 1997 letter as being no longer in question (Trovato 1997C Enclosure1).

3.13 2918 CASTILER - GRIDFLO: Index for Selecting Brine Pocket Volume

17

The parameter CASTILER - GRIDFLO is an index for selecting the brine pocket volume in theBRAGFLO code. This parameter was treated as a sampled variable in the CCA PA ranging from1 to 32. A CASTILER - GRIDFLO value of 1 selects the smallest volume pocket (pore volumeof 32,000 m3), and a value of 32 selects the largest volume pocket (pore volume of 160,000 m3).An explanation of DOE’s rationale for this parameter is presented in Section PD-1.9 of the SAReport (1998). In its initial review, the Agency did not find this parameter to be adequatelysupported. This parameter is potentially significant in computing brine releases associated withCastile brine pockets and was identified in the Agency’s March 19, 1997 letter for furtherevaluation (Trovato 1997A Enclosure 2).

One of the analyses initially planned by the Agency was to test model sensitivity to the extremevalues of the CCA brine pocket volume range. This analysis is described in Section PD-1.9 ofthe SA Report (1998). Simultaneously, the Agency tested model sensitivity to much largervolumes (3,400,000 m3 to 17,000,000 m3) as described in Section PD-1.10 of the SA Report(1998). When the sensitivity to the larger volumes was found to be only 1% (see SA Report 1998Table 3.1-1), further evaluation of the smaller CCA range of brine pocket volumes wasdiscontinued due to a low sensitivity and the parameter CASTILER - GRIDFLO was identifiedin the Agency’s April 25, 1997 letter as being no longer in question (Trovato 1997C Enclosure1). Additional discussion of the Castile brine pocket volume is presented in Section 4.5 of thisreport.

18

4.0 PARAMETERS WITH DIFFERENT VALUES OR RANGES

Parameters for which the Agency reviewed the supporting information and found that theinformation in the record supported a value or range of values different from those selected byDOE are addressed in this section. These parameters are listed in Enclosure 3 of the Agency’sMarch 19, 1997 letter (Trovato 1997A). Each of the following subsection headings identifies theparameter’s number, Material ID, and Parameter ID, followed by a brief description of theparameter.

4.1 3493 GLOBAL - PBRINE: Probability of Encountering Pressurized Brine

This parameter was used in the CCDFGF code for determining the frequency with which apressurized Castile brine pocket would be encountered by an intrusion borehole. This parameterwas treated as a constant in the CCA PA equal to 0.08, based on the results of a geostatisticalanalysis of pressurized brine pocket encounters in the vicinity of the WIPP Site (see WPO #40199 p. 36). Additional quantitative information from the results of time domainelectromagnetic (TDEM) geophysical surveys at the Site did not appear to have been used andsupported a higher value for this parameter than was used by DOE in the CCA PA (The EarthTechnology Corporation 1988 and A-93-02, Reference #563, Volume 3, Chapter 6, Table 6.0-3).Although the authors of the geophysical report did not provide an estimated value for thisparameter, in a subsequent analysis of the TDEM data by SNL, the value of this parameter wasestimated to range from 0.10 to 0.55 with an expected value of 0.25, depending on the depth ofthe pocket in the Castile (see WPO # 39121 p. 2). The DOE believes that the value of thisparameter is lower than that inferred from the geophysical results because of the reducedprobability of hitting subvertical fractures containing the brine with vertical boreholes. However,this argument was not used in a quantitative manner to support the lower parameter value. Thisparameter has a potentially significant influence on the amount of brine that could enter therepository and affect direct brine releases and spallings releases during drilling. It was thereforelisted in Enclosure 3 of the Agency’s March 19, 1997 letter as requiring further evaluation(Trovato 1997A).

The sensitivity of the PA model to changes in the probability of encountering pressurized brinein the Castile was studied in the Agency’s SA using a low probability value of 0.01 and a highvalue of 0.60 (see SA Report 1998 Appendix PD Section PD-5.1). The Agency assigned arelatively broad range of values to this parameter to reflect its high degree of uncertainty.

The low value was selected by the Agency to represent a reasonable minimum that bounded the1992 PA and the geostatistical study. The Agency believes that the DOE’s geostatisticalanalysis, the aforementioned subvertical fracture hypothesis, and the low end of the rangeidentified in WPO # 39121 have some merit, and while there is reasonable evidence that one ormore brine pockets may be present beneath the site based on the TDEM data, it is not certain(i.e. a probability of 1) that a borehole penetrating such a geophysical anomaly would actuallyencounter pressurized brine with sufficient productivity to affect waste isolation. The Agencytherefore believes that the true probability of encountering pressurized brine beneath the WIPP

19

waste area could be low, and considers 0.01 to be a reasonable lower bound.

The high value was developed in consideration of the TDEM data and represents a reasonablemaximum in view of the aforementioned alternative interpretations that have been based on thesedata (see WPO # 40199 p. 36 and WPO # 39121 p. 2). Because of the uncertainty in thisparameter, as discussed in the previous paragraph, and consistent with the Agency’s approach inmaking the low value (0.01) lower than the results obtained in DOE’s analysis (0.08), theAgency selected a high value (0.60) that is higher than the highest value that has been interpretedfrom the TDEM data (0.55). The Agency has also considered the possibility that the WIPP-12brine reservoir may underlie 100% of the site and therefore the probability of encounteringpressurized brine would be 100%. This consideration is based on the assumption that the WIPP-12 reservoir is cylindrical in shape, which the Agency considers unlikely, and on the assumptionthat if present beneath the site, the reservoir is certain to yield sufficient brine to affect wasteisolation, which the Agency considers unreasonable. Although the Agency agrees that part of theWIPP-12 reservoir may underlie part of the site, the TDEM data do not support speculation of a100% probability of encounter. In view of the lack of support from the TDEM data and the otherconcerns expressed above, the Agency did not believe that an upper bound value higher than0.60 was reasonable. The Agency also considered the possibility that the probability ofencountering a brine reservoir would be zero. This possibility was also rejected based on theTDEM data which indicate the clear potential for a brine reservoir to lie beneath the site.

The sensitivity analysis showed that the CCDFs upon which regulatory compliance is basedwere not sensitive to changes ranging from the low to high value of this parameter (see SAReport 1998 Table 3.5-1). Despite a demonstrated low sensitivity, the Agency continued todisagree with DOE’s technical approach for this parameter and based on the concerns expressedin public comments, the Agency required DOE to modify it in the PAVT by treating it as asampled variable with a uniform distribution and a range of 0.01 to 0.60 (see Trovato 1997CEnclosure 2). A uniform distribution was mandated because the range of this parameter spansslightly more than an order of magnitude and the use of a uniform distribution willconservatively bias the sampling toward the high end. The range is the same as used in the SAand is based on the same rationale as described above for the SA. These changes increased themedian value of this parameter from 0.08 to 0.305 and are expected to conservatively increasethe frequency of encountering a Castile brine pocket in the PAVT.

4.2 2254 BOREHOLE - TAUFAIL: Waste Shear Resistance

The waste shear resistance is a measure of the resistance of the waste to erosion by movingborehole fluids at the waste-borehole interface. This parameter was used in the CUTTINGS_Scode for calculating cavings releases. It was treated as a sampled value in the CCA PA with auniform distribution and a range of 0.05 to 10.0 Pa. This range of values was derived by DOEfrom literature studies of erosion in San Francisco Bay mud and consideration of the meanparticle size of the WIPP waste (see WPO # 40521 p. 18 and CCA, Volume XI, Appendix PARp. PAR-117 [Docket: A-93-02, II-G-1]). This parameter represents the threshold value of fluidshear stress required to sustain general erosion of the borehole wall. Its value is considerably

20

lower than and is not the same as the normal soil shear strength.

A similar parameter was independently assigned a quite different value by DOE in the spallingsrelease calculations. That parameter was called the waste cementation strength. It is a measure ofthe resistance of the waste to erosion by moving pore gas and was assigned a constant value of700 psi (6,895 Pa) (parameter BLOWOUT - CEMENT; see Section 3.1). Because bothparameters are measures of resistance to erosion by moving fluids, the Agency considered thatthe basis for their assigned values should be correlated. The Agency’s disposition of parameterBLOWOUT - CEMENT is described in Section 3.1. The parameter BOREHOLE - TAUFAILhas a significant influence on cavings releases and was therefore listed in Enclosure 3 of theAgency’s March 19, 1997 letter as requiring further evaluation (Trovato 1997A).

The sensitivity of the PA model to changes in the waste shear resistance was studied in theAgency’s SA using a low value of 0.01 Pa and a high value of 10 Pa (see SA Report 1998Appendix PD Section PD-3.2). The low value was of most interest to the Agency because aweaker material would result in greater cavings releases. The low value was originally deemedunsupported but DOE provided additional erosional studies that suggested the minimum shearstress would actually be 0.1 Pa or greater. EPA accepted the 0.05 Pa as being acceptablebecause it is conservative. The high value was set at 10 Pa and is the same as the high end of theCCA range. The results of this analysis showed a sensitivity of 1413% to the change in thisparameter, which the Agency considered to be significant (see SA Report 1998 Table 3.3-1).Because of its demonstrated sensitivity, the Agency required DOE to modify this parameter inthe PAVT based on the results of the particle size distribution expert elicitation (see Trovato1997C Enclosure 2 and Section 3.2 of this report).

DOE’s proposal for estimating the value of parameter BOREHOLE - TAUFAIL based on theparticle size distributions determined by the Expert Elicitation Panel (EEP) is summarized in aJune 27, 1997 memorandum (WPO # 46646). The basis for convening such a panel for thisparameter is discussed in Section 3.2 of this report. Based on the EEP results, the volume-averaged mean particle size for fully degraded waste was determined by DOE to range from alow of 0.1 cm to a high of 10 cm (assuming no cementation) or room-size (assumingcementation) (see WPO # 46936 and Section 3.2 of this report). The minimum value of 0.1 cm ismuch greater than the 40 microns used in the CCA spallings calculations. For conservatism,DOE based its estimate on the lower range of 0.1 to 10 cm and the minimum waste particledensity of 2.5 gm/cm3. DOE’s approach used the Shield’s parameter as mandated by the Agency(Trovato 1997C Enclosure 2), which relies on a measure of the central point of a population ofparticles of various sizes to determine the critical shear stress for an erodible, cohesionlesssediment bed (Simon and Senturk 1992). Based on this approach, the DOE calculated criticalshear stresses ranging from 0.64 Pa to 77 Pa. This range is higher than that used in the CCA PAand would lead to lower cavings releases. For conservatism, DOE proposed and the Agencyaccepted that the waste shear resistance in the PAVT be sampled from a log-uniform distributionwith a range of 0.05 to 77 Pa. The log-uniform distribution was selected to provide equalweighting over the three orders of magnitude in the range. The low value was equal to the lowend of the range used in the CCA, which DOE supported with additional information. This

21

information indicated that very fine-grained materials are not cohesionless as assumed in theAgency’s aforementioned Shields Parameter calculation. The information also showed that alower bound of the critical shear stress for fine-grained, cohesive sediments is on the order of0.05 Pa (Parthenaides and Paaswell 1970). The high end of the range was considered appropriatefor cohesionless particles and was retained based on the EEP results.

4.3 3184 BH_SAND - PRMX_LOG: Long-Term Borehole Permeability

The long-term borehole permeability is the permeability assigned to an intrusion borehole whenit is assumed to have been filled with casing corrosion products, plug degradation products, andnative materials sloughed from overlying formations. DOE estimated the value of this parameterfrom a literature review using silty sand as a surrogate material (Freeze and Cherry 1979 Table2.2). Because of its uncertainty, it was treated in the CCA as a sampled variable with a uniformdistribution and a range of -14 to -11 log m2. The Agency did not agree with the lower bound ofthis range and believes that the value may be may be closer to that of a borehole plug. TheAgency was also not satisfied that DOE’s range adequately captured the uncertainty of thisparameter. This parameter has a potentially significant influence on long-term releases fromintrusion boreholes and was listed in Enclosure 3 of the Agency’s March 19, 1997 letter asrequiring further evaluation (Trovato 1997A).

The sensitivity of the PA model to changes in the long term borehole permeability was studied inthe Agency’s SA using a low value of -17 log m2 and a high value of -11 log m2 (see SA Report1997 Appendix PD Section PD-1.5). The low value was of most interest to the Agency because aless permeable material could result in greater gas pressure buildup in the repository withconsequent increases in brine and spallings releases. It was selected based on the permeability ofpartially degraded concrete to reflect the possibility that the borehole plugs may not degrade asexpected by DOE, but may instead remain as low-permeability barriers over the regulatory timeframe. Information on the permeability of an intact borehole plug in salt obtained by DOE fromfield measurements at the WIPP Site and supporting laboratory tests was used by the Agency indetermining this lower bound value (Christensen and Hunter 1980 Figure 4).

The upper bound of the permeability range was not changed in the sensitivity analysis becausefurther reduction in this value would result in lower gas pressures and lower long-term releases.The analysis showed a 330% average change in the performance measures, which the Agencyconsidered to be significant (see SA Report 1998 Table 3.1-1). Because of its demonstratedsensitivity, the Agency required DOE to treat this parameter (and its Y- and Z-directioncounterparts) in the PAVT as a sampled variable with a uniform distribution and a range of -16.3log m2 to -11 log m2 (see Trovato 1997B Enclosure 2 and Kruger 1997). The uniformdistribution is the same as used in the CCA and was not changed. A uniform distribution wasaccepted for this logarithmic parameter because, based on present knowledge, the Agencybelieves that all values across the range of several orders of magnitude are equally likely. Thelower bound of -16.3 log m2 is the permeability assigned by DOE to an intact borehole plug (seeSection 5.17 of this report). It was used by the Agency as the lower bound of this range becauseit approximates conditions that would occur if the plug degradation assumed by DOE in the PA

22

did not occur. By selecting a low value of the sampled range based on the permeability of anundegraded plug, the analysis has the same type of impact as sampling an extended distributionof borehole plug lifetimes. Both analyses would include realizations with low plugpermeabilities that constrain gas release through the borehole over long periods of time. For thesame reasons as in the sensitivity analysis, the upper bound was allowed to remain the same asused in the CCA. Reducing the lower bound of this parameter was expected to result inconservatively higher repository gas pressures and higher radionuclide releases duringinadvertent intrusions.

4.4 2918 CASTILER - VOLUME: Castile Brine Pocket Volume

The volume of a Castile brine pocket is a potentially important parameter related to the volumeof brine that may be available to enter the repository, the rate of gas pressure buildup from wastecorrosion and degradation, and the brine and spallings releases to the accessible environment.The parameter CASTILER - VOLUME was listed among the input parameters to the PA modeland because considerably larger values for this parameter were available in DOE’s supportingdata, it was included in Enclosure 3 of the Agency’s March 19, 1997 letter as requiringadditional evaluation (Trovato 1997A). Upon further investigation, the Agency found that thisparameter was not used by DOE in the 1996 CCA. Instead, the brine pocket volume wassampled from a predetermined range using the index parameter CASTILER - GRIDFLO and aporosity correction term (see SA Report 1998 Appendix PD Sections PD-1.9 and PD-1.10).Because it was not used in the CCA PA model, the parameter CASTILER - VOLUME wasidentified in the Agency’s letter of April 25, 1997 as being no longer in question (Trovato 1997CEnclosure 1).

As an alternative to this parameter and to achieve the same objective of incorporating largerbrine pocket volumes, the Agency required DOE to modify the porosity correction terms in thePAVT. Because of the link between brine pocket volume and rock compressibility indetermining the volume of brine that can be released from a brine pocket, and the simultaneoustreatment of both parameters in the Agency’s sensitivity analysis, this modification is discussedalong with Castile brine pocket rock compressibility in Section 4.5 below.

4.5 61 CASTILER - COMP_RCK: Castile Brine Pocket Rock Compressibility

The Castile brine pocket rock compressibility is used along with other parameters in the PAmodel to calculate the volume of brine released from a Castile brine pocket. DOE estimated thevalue of this parameter based on a combination of field and literature data (see CCA Docket: A-93-02, II-G-1, Volume XI, Appendix PAR p. PAR-107 and WPO # 31084 p. 1). Because of itsuncertainty, this parameter was treated as a sampled variable with a triangular distribution, arange of 5 E-12 Pa-1 to 1 E-8 Pa-1, and a mode of 1 E-10 Pa-1. Analysis by DOE subsequent to theCCA PA calculations suggested a different range, one that was more closely tied to field testsconducted in borehole WIPP-12 in a Castile brine reservoir on the WIPP Site (see WPO # 41887pp. 1-3 and WPO # 44699 p. 1). Because of its relationship to the volume released from a brinereservoir, this parameter has a potentially significant influence on the volume of brine that may

23

be available to enter the repository, the rate of gas pressure buildup from waste corrosion anddegradation, and the brine and spallings releases to the accessible environment. This parameterwas therefore listed in Enclosure 3 of the Agency’s March 19, 1997 letter as requiring furtherevaluation (Trovato 1997A).

The sensitivity of the PA model to changes in the Castile brine pocket rock compressibility wasstudied in two different analyses conducted by the Agency. In the first of these sensitivityanalyses, only the rock compressibility was varied to determine the sensitivity to variations inthat parameter alone (see SA Report 1998 Appendix PD Section PD-1.8). In performing this firstanalysis, the compressibility was set at high and low values corresponding to the extremes of therange determined by DOE. During this analysis, the Castile brine pocket porosity was 0.0087,the parameter GRIDFLO was set at a median value of 16 corresponding to the CCA medianbrine pocket pore volume of 1.28 E+05 m3, and all other sampled PA inputs were similarly fixedat their median values. The reservoir productivity ratio (pore volume times pore compressibility,or alternatively, pore volume times rock compressibility divided by porosity) corresponding tothese values ranges from 7.4 E-05 to 1.5 E-01 m3/Pa and is an estimate of the brine volume thatthe brine pocket is capable of releasing to an intrusion borehole per unit change in brine pocketpressure. These productivity ratios bracket the expected productivity ratio of approximately 4 E-02 m3/Pa estimated for the brine reservoir encountered at WIPP-12 (based on a total pore volumeof 2.7 E+06 m3 [17 E+06 bbl] and a pore compressibility of 1.45 E-08/Pa [100 E-06/psi]; seeWPO # 42085 p. H-53). This sensitivity analysis is therefore also considered by the Agency toprovide a test of the sensitivity of PA model results to changes in the volume of brine released bya Castile brine pocket. The Agency considers the WIPP-12 reservoir to provide an appropriatelyconservative reference value for reservoir properties because of its relatively large size and itsproximity to the WIPP Site. Although the compressibility and therefore the reservoirproductivity ratio variation ranged over nearly four orders of magnitude in the sensitivityanalysis, the average change in the performance measures was only 2% and was not consideredto be significant (see SA Report 1998 Table 3.1-1).

In the second of these sensitivity analyses, the compressibility was varied along with the porositycorrection term to effect a significant increase in the brine pocket volume (see SA Report 1998Appendix PD Section PD-1.10). In this analysis the porosity correction term was assigned a lowvalue of 0.1848, corresponding to a brine pocket pore volume of 3.4 E+06 m3, and a high valueof 0.924, corresponding to a brine pocket pore volume of 1.7 E+07 m3 (see SA Report 1998Appendix PD Section PD-1.10 for additional information on this analysis and an explanation ofthe relationship between this correction term and the brine pocket pore volume). This brinepocket volume range was suggested by DOE based on an analysis of the WIPP-12 brine pocketand is considered by the Agency to provide a reasonable estimate of the possible range ofvolumes that may be present at the site (see WPO # 41887). In this second sensitivity analysisthe productivity ratio ranged from 7.8 E-03 to 2.0 E-01 m3/Pa and also bracketed theaforementioned WIPP-12 productivity ratio. Again, the average change in the performancemeasures was low, only 1%, and was not considered significant.

Even though the sensitivity of the PA model to changes in rock compressibility and brine pocket

24

volume was found to be low, the Agency was not satisfied with DOE’s technical justification forthe brine pocket rock compressibility and pore volume range used in the CCA PA and because ofnumerous public comments the Agency required DOE to change those ranges in the PAVT(Trovato 1997B Enclosure 2). In the Agency’s letter, the DOE was required to treat the rockcompressibility as a sampled variable with a triangular distribution, a revised range of 2 E-11 Pa-

1 to 1 E-10 Pa-1, and a revised mode of 4 E-11 Pa-1. This revised range was estimated by DOEand accepted by the Agency based on an analysis of the Castile brine pocket encountered inborehole WIPP-12 (WPO # 41887). As previously mentioned, the Agency considered the WIPP-12 reservoir to provide appropriately conservative reference values for Castile brine reservoirproperties and therefore considered this analysis to be appropriate and the results preferable todata taken from other formations at WIPP or from non-WIPP sources reported in the literature.The form of the distribution was not changed and remained the same as in the CCA. The Agencyconcurred with the form of the distribution selected by DOE because, although a simple formwas warranted because little was known about the distribution’s specific shape, the Agencyagreed that midrange values were more likely than the extreme bounding values.

Although not documented in the aforementioned Agency letter, the Agency also requested theDOE to change the porosity correction term in the PAVT to a sampled variable with a triangulardistribution, a range of 0.1848 to 0.9240, and a mode of 0.3696. A new parameter (8000CASTILER - POR_BPKT) was created in the PA model to allow the porosity correction term tobe sampled. The triangular distribution reflects the Agency’s belief that an intermediate value ismore likely that either extreme value, yet it retains a simple shape that reflects the uncertaintyabout the distribution’s specific shape. The mode is equivalent to a brine pocket pore volume of6.8 E+06 m3 and is equal to the average of the range of WIPP-12 brine pocket volumes estimatedby Popielak et al. (1.7 E+06 to 8.7 E+07 bbl; see WPO # 42085 p. H-60). The range wasdetermined from a reanalysis of the WIPP-12 data. It is higher than the range of Popielak et al.and is the same as used in the aforementioned second sensitivity analysis (see WPO # 41887).The productivity ratio for the PAVT has the same range as for the second sensitivity analysis(7.8 E-03 to 2.0 E-01 m3/Pa) and brackets the WIPP-12 productivity ratio of 4 E-02 m3/Pa.These changes have the effect of modeling the characteristics of Castile brine pockets in thePAVT after those of the WIPP-12 brine reservoir. The parameter changes in the PAVT reducedthe rock compressibility range and mode to more appropriate values and increased the brinereservoir volume. The reduced compressibility would tend to reduce the reservoir’s capability torelease brine, while the increased volume would tend to increase that capability. Because thevolume increase was substantially greater than the compressibility reduction, the net effect isexpected to conservatively increase the volume of brine available to flow from the reservoir.

25

5.0 PARAMETERS WITH QUESTIONABLE VALUES OR RANGES

Parameters for which the Agency has reviewed the supporting information and has questionsabout the values selected are addressed in this section. These parameters are listed in Enclosure 4of the Agency’s March 19, 1997 letter (Trovato 1997A). Each of the following subsectionheadings identifies the parameter’s number, Material ID and Parameter ID, followed by a briefdescription of the parameter.

5.1 27 BOREHOLE - DOMEGA: Drill String Angular Velocity

This parameter is the rotational velocity of the drill string in an intrusion borehole. It is used incalculating cavings release volumes in the CUTTINGS_S code. This parameter is treated as aconstant in the CCA and is equal to the median of a constructed cumulative distribution ofrotational velocities used in current practice when drilling through salt (7.8 radians per second;see WPO # 31512 p. 2 and WPO # 37765 p. 49). In its parameter review, the Agency found thatthe range of angular velocities used in salt was large, ranging from 4.2 to 23 radians per second(A-93-02, II-G-1, Volume V, Appendix Cuttings, Section 2.3). Because of the potentialimportance of this parameter in calculating cavings releases and the magnitude of the range, theAgency questioned the use of a single value rather than a range in the CCA and listed thisparameter in Enclosure 4 of the Agency’s March 19, 1997 letter as requiring further evaluation(Trovato 1997A).

The sensitivity of the PA model to drill string angular velocity was tested by the Agency byassigning low and high values equal to the extremes of the aforementioned range identified byDOE (see A-93-02, II-G-1, Volume V, Appendix Cuttings, Section 2.3 and SA Report 1998Appendix PD Section PD-3.1). The results of the analysis showed a 60% change in the cavingsreleases, which the Agency found to be potentially significant. In view of this demonstratedsensitivity and the potential importance of cavings releases in meeting the Agency’s regulatorycriteria, the DOE was required by the Agency’s April 25, 1997 letter to treat the drill stringangular velocity as a sampled variable in the PAVT with a constructed cumulative distribution, aminimum of 4.2 radians/second, a maximum of 23 radians/second, and a median of 7.77radians/second (Trovato 1997C Enclosure 2). A constructed cumulative distribution is adistribution based directly on original data. It was selected because sufficient data were availableto support this approach. This range is the same as used in the SA and was consideredappropriate by the Agency because, as described above, it is based on an acceptable study ofcurrent drilling practices in salt. These changes are expected to conservatively increase the rangeof cavings releases in the PAVT.

5.2 64 CASTILER - POROSITY: Castile Brine Pocket Porosity

The effective porosity of a Castile brine pocket is one of the parameters used in the BRAGFLOcode to calculate the brine pocket volume. Although this parameter is identified in the PAdatabase as a sampled variable with a Student T distribution, a minimum of 0.002, a maximumof 0.016, and a median of 0.0087, it is treated as a constant equal to the median value of 0.0087

26

in the CCA. These values are based on the results of three laboratory measurements of effectiveporosity performed on intact pieces of Castile anhydrite core (see WPO # 42085). Thedocumentation further stated that the porosity value used in the model was calculated bymultiplying the value of this parameter by the number of brine pockets, and dividing that productby 5 (see WPO # 40434). Because the purpose of this calculation was not understood and itsjustification was not available at the time of the initial review, the Agency included thisparameter in Enclosure 4 of its March 19, 1997 letter as requiring further evaluation (Trovato1997A).

In subsequent reviews, the Agency learned that the aforementioned calculation was associatedwith DOE’s assumption that brine pockets occur beneath the WIPP Site in five discrete sizeswith constant pore volume increments of 32,000 m3 and total pore volumes ranging from 32,000to 160,000 m3 (see the SA Report 1998 Appendix PD Section PD-9 for additional discussion ofthis approach). Although at first confusing , the Agency found this approach to be internallyconsistent and acceptable. This parameter was therefore identified in the Agency’s letter of April25, 1997 as being no longer in question (Trovato 1997C Enclosure 1).

5.3 66 CASTILER - PRESSURE: Castile Brine Pocket Pore Pressure

This parameter is the undisturbed brine pressure in a Castile brine pocket before intrusion by anexploration borehole. It was estimated by DOE based on available information on pressures inbrine pockets encountered in the Delaware Basin (see CCA Docket: A-93-02, II-G-1, VolumeXI, Appendix PAR p. PAR-101, WPO # 37148 p. 1, and WPO # 37973 pp. 1-3). Because of itsrelatively high uncertainty, this parameter was treated as a sampled variable in the CCA with atriangular distribution, a range of 1.11 E+07 Pa to 1.70 E+07 Pa, and a median of 1.27 E+07 Pa.Although the basis for selecting the range of pressures used in the CCA was well documented,the Agency found that the form of the distribution was insufficiently explained. Because of thepotential importance of this parameter in calculating the brine volume in the repository and theAgency’s questions regarding the distribution used in the CCA, it was listed in Enclosure 4 ofthe Agency’s March 19, 1997 letter as requiring further evaluation (Trovato 1997A).

The sensitivity of the PA model to the Castile brine pocket pore pressure was tested by theAgency by assigning low and high values equal to the extremes of the aforementioned rangeidentified by DOE (see SA Report 1998 Appendix PD Section PD-1.11). The results of theanalysis showed no model sensitivity (see SA Report 1998 Table 3.1-1). In view of this lack ofsensitivity, the DOE was informed in the Agency’s April 25, 1997 letter that this parameter wasno longer in question (Trovato 1997C Enclosure 1).

5.4 259 PAN_SEAL - PRMX_LOG: Panel Seal Permeability

This parameter is the permeability of the concrete and crushed salt seals used to isolate the wastepanels after they have been filled. The panel seal permeability was assigned by DOE to be thesame as that of the DRZ. Although the seal permeability would be expected to change over timeas the concrete degraded and the halite was compressed by creep closure, it was treated as a

27

constant in the CCA equal to -15 log m2. The uncertainty in this parameter led the Agency toquestion the DOE’s approach. Because of the potential importance of this parameter inrestricting brine movement within the repository and the Agency’s questions regarding the valueused in the CCA, it was listed in Enclosure 4 of the Agency’s March 19, 1997 letter as requiringfurther evaluation (Trovato 1997A).

The sensitivity of the PA model to the panel seal permeability was tested by the Agency bydetermining the impact of low and high parameter values on PA performance measures. Theeffect of a reduction in the CCA value was of primary interest to the Agency because it wouldrestrict gas flow within the repository. The effect of an increase in the CCA value was also ofinterest because it would allow brine to move more freely within the repository. In performingthe SA, the Agency used a low value of -21 log m2 and a high value of -13 log m2. Selecting alow value that is six orders of magnitude less than the CCA value and a high value that is twoorders of magnitude greater than the CCA value was considered to provide an adequate analysisof sensitivity (see SA Report 1998 Section 1.14). The results of the analysis showed no modelsensitivity (see SA Report 1998 Table 3.1-1). In view of this lack of sensitivity, the DOE wasinformed in the Agency’s April 17, 1997 letter that this parameter was no longer in question(Trovato 1997B Enclosure 1).

5.5 528 S_ANH_AB - POROSITY: Effective Porosity of Anhydrite Beds A and B

This parameter is the initial effective porosity of anhydrite beds A and B. Its value wasdetermined by DOE as the mean porosity of 16 field samples (see WPO # 34860 p. 2). Thisparameter was treated as a constant in the CCA equal to 0.011. Because of its potentialimportance in calculating actinide transport rates to the accessible environment, the Agencyquestioned the sensitivity of the PA code to the extreme porosity values found in the fieldsamples. Because of this uncertainty, the Agency listed this parameter in Enclosure 4 of theAgency’s March 19, 1997 letter as requiring further evaluation (Trovato 1997A).

The sensitivity of the PA model to changes in the anhydrite effective porosity value was tested aspart of the Agency’s sensitivity analysis. In performing this analysis, the effective porosity of allanhydrite beds was assigned a low value of 0.006 and a high value of 0.017 based on the extremeporosity values measured in the aforementioned field samples. (see SA Report 1998 AppendixPD Section PD-1.2). The results of the analysis showed no model sensitivity (see SA Report1998 Table 3.1-1). In view of this lack of sensitivity, the DOE was informed in the Agency’sApril 17, 1997 letter that this parameter was no longer in question (Trovato 1997B Enclosure 1).

5.6 567 S_MB138 - POROSITY: Effective Porosity of Anhydrite Marker Bed 138

This parameter is the initial effective porosity of anhydrite Marker Bed 138. Its was assigned thesame constant value by DOE as the effective porosity of anhydrite beds A and B discussed inSection 5.5 above, and consequently carried the same Agency concerns. Because of theseconcerns, the Agency listed this parameter in Enclosure 4 of the Agency’s March 19, 1997 letteras requiring further evaluation (Trovato 1997A). The sensitivity of the PA model to changes in

28

the value of this parameter was tested by the Agency in the same analysis as the effectiveporosity for beds A and B. The same low and high values were assigned as described in Section5.5, and no model sensitivity was found (see SA Report 1998 Table 3.1-1). In view of this lackof sensitivity, the DOE was informed in the Agency’s April 17, 1997 letter that this parameterwas no longer in question (Trovato 1997B Enclosure 1).

5.7 588 S_MB139 - POROSITY: Effective Porosity of Anhydrite Marker Bed 139

This parameter is the initial effective porosity of anhydrite Marker Bed 139. Its was assigned thesame constant value by DOE as the effective porosities of anhydrite beds A and B and MarkerBed 138 discussed in Sections 5.5 and 5.6 above, and consequently carried the same Agencyconcerns. Because of these concerns, the Agency listed this parameter in Enclosure 4 of theAgency’s March 19, 1997 letter as requiring further evaluation (Trovato 1997A). The sensitivityof the PA model to changes in the value of this parameter was tested by the Agency in the sameanalysis as the effective porosity for the other anhydrite beds. The same low and high valueswere assigned as described in Section 5.5, and no model sensitivity was found (see SA Report1998 Table 3.1-1). In view of this lack of sensitivity, the DOE was informed in the Agency’sApril 17, 1997 letter that this parameter was no longer in question (Trovato 1997B Enclosure 1).

5.8 651 WAS_AREA - ABSROUGH: Waste Area Absolute Roughness

This parameter is a measure of the roughness of the wall of an intrusion borehole where it passesthrough the waste. It was listed by DOE as an input to the CUTTINGS_S model for use incalculating cavings releases. Because the value of this parameter was poorly documented and thecavings volume was an important contributor to overall releases, it was included in Enclosure 4of the Agency’s March 19, 1997 letter as requiring further evaluation (Trovato 1997A). Uponfurther investigation, it was found that this parameter was not used by DOE in the 1996 CCA PAmodel to set the absolute roughness of the waste and had no effect on the model results. Thisparameter was therefore identified in the Agency’s letter of April 25, 1997 as being no longer inquestion (Trovato 1997C Enclosure 1).

5.9 653 WAS_AREA - COMP_RCK: Waste Area Rock Compressibility

This parameter is the compressibility of the waste material in the repository. DOE assigned azero value to this parameter in the CCA and noted that the compressibility of the waste isaddressed in the porosity surface room closure calculations (see CCA Docket: A-93-02, II-G-1,Volume I, Chapter 6, p. 6-101). Because this parameter was not well documented in the DOE’sdata base and the Agency questioned the use of a zero value by DOE, it was included inEnclosure 4 of the Agency’s March 19, 1997 letter as requiring further evaluation (Trovato1997A).

The sensitivity of the PA model to changes in the waste area rock compressibility was studied inthe Agency’s SA. A range for waste compressibility was calculated by the Agency based onwaste strength information presented in the 1992 Draft Performance Assessment (SNL 1992 Vol

29

3 p. 2-69). Upon further investigation, it was found that although this parameter was retained byDOE in the 1996 CCA PA model, it was not actually used in the calculations and its value hadno effect on the model results. This parameter was therefore identified in the Agency’s letter ofApril 25, 1997 as being no longer in question (Trovato 1997C Enclosure 1).

5.10 1992 WAS_AREA - DIRNCCHW: Bulk Density of Iron Containers for CH Waste

This parameter is the average bulk density of iron containers for contact-handled waste at WIPP.It was used in the CCA to calculate the contribution of the iron containers themselves to the totalmass of iron available for corrosion in the repository. The value of this parameter wasdetermined by DOE from information in the Baseline Inventory Report (BIR), which documentsthe physical characteristics of the waste containers expected to be received by the WIPP (WPO #32328). The total mass of iron in the repository is used along with brine inflow and otherparameters to calculate gas generation rates and repository gas pressures, which have been foundto be important drivers for brine and spallings releases. Because this parameter is potentiallyimportant in repository performance calculations and adequate documentation was not found inthe Agency’s initial review, it was included in Enclosure 4 of the Agency’s March 19, 1997letter as requiring further evaluation (Trovato 1997A).

Upon completing a comprehensive evaluation of the BIR (EPA 1998a), the Agency confirmedthat the necessary supporting information for this parameter was appropriately documented andthat the related uncertainty was low. This parameter was therefore identified in the Agency’sletter of April 17, 1997 as being no longer in question (Trovato 1997B Enclosure 1).

5.11 1993 WAS_AREA - DIRNCRHW: Bulk Density of Iron Containers in RH Waste

This parameter is the average bulk density of iron containers for remote-handled waste at WIPP.It was used in the CCA to calculate the contribution of the iron containers themselves to the totalmass of iron available for corrosion in the repository. The value of this parameter wasdetermined by DOE from information in the Baseline Inventory Report (BIR), as discussed inSection 5.10 for the parallel parameter for contact-handled waste (WPO # 32328). Because thisparameter is potentially important in repository performance calculations and adequatedocumentation was not found in the Agency’s initial review, it was included in Enclosure 4 ofthe Agency’s March 19, 1997 letter as requiring further evaluation (Trovato 1997A).

Upon completing a comprehensive evaluation of the BIR (EPA 1998a), the Agency confirmedthat the necessary supporting information for this parameter was appropriately documented andthat the related uncertainty was low. This parameter was therefore identified in the Agency’sletter of April 17, 1997 as being no longer in question (Trovato 1997B Enclosure 1).

5.12 2040 WAS_AREA - DIRNCHW: Average Density of Iron-Based Material in CH Waste

This parameter is the average density of iron-based material in contact-handled waste at WIPP.Its value was determined by DOE from information in the Baseline Inventory Report (BIR),

30

which documents the physical characteristics of the wastes expected to be received at WIPP(WPO # 32328). It was used by DOE to determine the mass of iron-based material in the waste.Because this parameter appeared potentially important in repository performance calculationsand adequate documentation was not found in the Agency’s initial review, it was included inEnclosure 4 of the Agency’s March 19, 1997 letter as requiring further evaluation (Trovato1997A).

The sensitivity of the PA model to changes in waste density was studied in five analysesperformed by the Agency. These studies were performed by determining the change in modelperformance measures resulting from varying the average density of iron-based waste materials,cellulosics, rubber, and plastics in the waste. The values of each of these densities were changedboth separately and collectively, and the results are described in the Agency’s SA Report (1997Section 3.1). The average bulk densities of iron containers was not varied in these analysesbecause their uncertainty is low. The average changes of the PA performance measures due tochanges in waste density parameters were found to be greater than for any other parameterstudied, and ranged from about 70,000% to 104,000%. The high sensitivity of the repositoryperformance measures to changes in waste density (or to the calculated mass of waste in therepository) was not unexpected because the SA was conducted assuming that all waste receivedwas either the maximum or the minimum density. However, it should be noted that becausewaste density was an input to the BRAGFLO code, the Agency’s performance measures were aseries of key outputs such as repository gas pressures and repository brine flows that arecalculated by BRAGFLO, rather than direct releases to the accessible environment that are notcalculated by BRAGFLO. The selection of performance measures for the SA is more fullydiscussed in Appendix PM of the SA Report (1997).

Because of the high sensitivity to changes in waste density, the approach taken by DOE indetermining density data for the CCA PA was reviewed by the Agency in detail (EPA 1998a). The average densities for all waste types were found to be appropriately calculated from datapresented in the BIR (WPO#32328), and account for the expected waste inventory. The Agencyconsidered the minimum and maximum densities presented by the generator sites in AppendixBIR, Appendix P (A-93-02, II-G-1, Volume III-V, Appendix BIR) of the CCA and determinedthat use of the lowest and highest minimum and maximum values in PA would not be a realisticrepresentation of the expected inventory. For example, although one generator site identified aminimum density of 0 for a parameter in a site-specific waste stream, it would be inappropriateto apply this value to all similar WIPP waste streams, or to use this as a representative minimumvalue for the parameter.

EPA examined the average values provided by generator sites on a site-specific waste streambasis and also on an average basis across all WIPP waste streams. The Agency found that use ofthe average value for waste densities provides a realistic and justifiable representation of theexpected waste inventory. The density parameters were therefore identified in the Agency’sletter of April 17, 1997 as being no longer in question (Trovato 1997B Enclosure 1).

5.13 2041 WAS_AREA - DCELLCHW: Average Density of Cellulosics in CH Waste

31

This parameter is the average density of cellulosics in contact-handled waste at WIPP. Its valuewas determined by DOE from information in the Baseline Inventory Report (BIR), whichdocuments the physical characteristics of the wastes expected to be received at WIPP (WPO #32328). It was used by DOE to determine the mass of cellulosic materials in the repository. Thisparameter was included along with other waste density parameters in Enclosure 4 of theAgency’s March 19, 1997 letter as requiring further evaluation (Trovato 1997A). Uponevaluating this parameter in detail, the Agency reached the same conclusions as discussed inSection 5.12 above. Although PA modeling results were found to be sensitive to its value, for thereasons given in Section 5.12 this parameter was identified to the DOE in the Agency’s letter ofApril 17, 1997 as being no longer in question (Trovato 1997B Enclosure 1).

5.14 2274 WAS_AREA - DCELLRHW: Average Density of Cellulosics in RH Waste

This parameter is the average density of cellulosics in remote-handled waste at WIPP. Its valuewas determined by DOE from information in the Baseline Inventory Report (BIR), whichdocuments the physical characteristics of the wastes expected to be received at WIPP (WPO #32328). It was used by DOE to determine the mass of cellulosic materials in the repository. Thisparameter was included along with other waste density parameters in Enclosure 4 of theAgency’s March 19, 1997 letter as requiring further evaluation (Trovato 1997A). Uponevaluating this parameter in detail, the Agency reached the same conclusions as discussed inSection 5.12 above. Although PA modeling results were found to be sensitive to its value, for thereasons given in Section 5.12 this parameter was identified to the DOE in the Agency’s letter ofApril 17, 1997 as being no longer in question (Trovato 1997B Enclosure 1).

5.15 2907 STEEL - CORRMCO2: Steel Corrosion Rate

This parameter is the rate of steel corrosion in a brine-saturated repository without CO2 presentin the brine. Use of this parameter assumes the presence and complete effectiveness of the MgObackfill in reacting with and removing the CO2 from solution. The steel corrosion rate wasestimated by DOE based on long-term anoxic steel corrosion experiments (see CCA Docket: A-93-02, II-G-1, Volume XI, Appendix PAR p. PAR-15). Because of its uncertainty, this parameterwas treated as a sampled variable in the CCA with a uniform distribution, a range from zero to1.59 E-14 m/s, and a median of 7.94 E-15 (see WPO #35162 p. 5 and WPO #35181 p. 1, both inWPO #30819). Because the Agency questioned both the lower and upper bounds for thisparameter, it was included along with other waste density parameters in Enclosure 4 of theAgency’s March 19, 1997 letter as requiring further evaluation (Trovato 1997A).

The sensitivity of the PA model to changes in the steel corrosion rate was studied in theAgency’s SA. Based on a review of the experimental corrosion data, the Agency agreed thatunder some circumstances the rate of corrosion could be very low but may not be zero. TheAgency’s objective in the sensitivity analysis was to determine the effect of a non-zero lowerbound to the corrosion rate, and accordingly increased the lower bound to 1 E-30 m/s to checkthe sensitivity to a small but nonzero value. The baseline and high values remained equal tothose used by DOE in the CCA. Although the results indicated that the performance measures

32

were not sensitive to this change in the lower bound (an average change of 7% was found; seeSA Report 1997 Table 3.1-1), the upper bound remained in question and was found in DOE’sexperimental results to be potentially sensitive to several aspects of the repository environment..

The effect of contact with salt on corrosion rates was evaluated in DOE tests. A DOE test inwhich steel was embedded in a bentonite-crushed salt mixture yielded a corrosion rateapproximately double the rate with no bentonite present (Telander and Westerman 1997 p. 6-39).However, a similar test conducted without bentonite present showed no rate increase (5 andWesterman 1995 p. 6-43). The WIPP design presented in the CCA does not include a bentonitebackfill and based on the foregoing experimental results, the Agency does not consider contactwith salt alone to significantly affect corrosion rates.

The Agency considers DOE’s use of long-term test results to be appropriate considering that thecorrosion rate parameter is intended to reflect repository behavior over 10,000 years. Isolatedlocal conditions that may result in higher short-term corrosion rates would have little impact onoverall, long-term repository conditions. DOE’s experiments were conducted under a pH of 9.5to 10, which represents the lowest pH that would occur in the repository in the presence of anMgO backfill that sequesters CO2 from solution (see Wang and Brush 1996 p. 5). Although alocal, lower pH environment could result in higher corrosion rates, based on the findings of theDOE’s independent Conceptual Models Peer Review Panel (Wilson et al. 1997B p. 12ff) and theAgency’s own review, the MgO backfill is expected to be effective in controlling pH.

An average corrosion rate of 2.25 E-14 m/s was determined by DOE from the average hydrogengas generation between the 12th and 24th months of a test in a sealed chamber where the steelspecimens were immersed in brine and over-pressured with nitrogen at a maximum end-of-testpressure of 17 atm. To obtain the maximum CCA value, this rate was adjusted downward by afactor of 70% based on another test of 38.5 months duration where a small amount of CO2 wasadded to the nitrogen cover gas (Telander and Westerman 1995). However, other experiments ofsix months duration conducted with a hydrogen cover gas indicated that the corrosion rate firstdecreased at overpressures from 2 to 70 atm and then increased at overpressures from 70 to 127atm (lithostatic pressure at the repository horizon is about 150 atm). An additional test with anitrogen cover gas produced a corrosion rate about twice as high as a companion test at 10 atm. Because the repository pressure may, under some circumstances, approach or exceed lithostaticand because the experimental corrosion rates increase at these higher pressures, in the Agency’sletter of April 17, 1997 (Trovato 1997B Enclosure 2) the DOE was required to double the upperbound in the PAVT to 3.17 E-14 m/s to reflect the effect of factors such as increased pressure onthe rate. A greater increase in the upper bound was not supported by available data. Highercorrosion rates are expected to have a conservative influence on repository performance becausethey result in increased rates of gas generation. At longer times, the effect of an increasedcorrosion rate is expected to be reduced because of limited brine availability.

The DOE’s experimental results also indicated that when aluminum is present the steel corrosionrate increased markedly during the first 13 months of a test but then remained constant for thenext 11 months (Telander and Westerman 1995). Telander and Westerman concluded that the

33

corrosion enhancement did not appear, on the basis of the limited tests performed, to be a long-lived effect. The Agency believes that this position is reasonable, but notes that the increasedrate used in the PAVT is reflective of such a possibility.

The Agency did not change the lower bound of zero used by DOE in the CCA. Although theAgency does not expect the rate to drop to identically zero, it could drop to a low value;however, the Agency’s sensitivity analysis showed that such a change would have no significantimpact on PA modeling results and would not be necessary in the PAVT. The Agency did notchange the uniform distribution used by DOE in the CCA because this parameter spans severalorders of magnitude and sampling from such a distribution conservatively favors highercorrosion rates.

5.16 3147 CONC_PLG - POROSITY: Borehole Plug Porosity

This parameter is the effective porosity assigned to a concrete plug expected to be placed bydrillers when sealing an intrusion borehole. DOE estimated borehole plug porosity fromengineering data on field-emplaced concrete structures. Concrete porosities were estimated torange from 25% to 40%, with a median value of 32% (CCA Docket: A-93-02, II-G-1, Volume XAppendix MASS Attachment 16-3 p. D-2 and Figure C-1). This parameter was treated as aconstant by DOE in the CCA PA equal to 0.32. The Agency questioned the use of a constantvalue for this parameter in light of the ranges provided for it in the literature and also in view ofits potential importance in determining transport rates through plugged boreholes. Because ofthese concerns, this parameter was included in Enclosure 4 of the Agency’s March 19, 1997letter as requiring further evaluation (Trovato 1997A).

The sensitivity of the PA model to changes in the borehole plug porosity was studied in theAgency’s SA. A range for this parameter of 0.24 to 0.40 was determined by the Agency based onthe aforementioned literature information. The results of the analysis showed no sensitivity ofthe performance measures to these changes, and this parameter was therefore identified in theAgency’s letter of April 17, 1997 as being no longer in question (Trovato 1997B Enclosure 1).

5.17 3185 CONC_PLG - PRMX_LOG: Borehole Plug Permeability

This parameter is the permeability assigned to a concrete plug expected to be placed by drillerswhen sealing an intrusion borehole and is effective for 200 years in the CCA PA calculations.DOE estimated borehole plug permeability through direct measurement of the permeability of aconcrete borehole plug at the WIPP Site (Christensen and Hunter 1980 Figure 4). This parameterwas treated as a constant in the CCA equal to -16.3 log m2 (see CCA Docket: A-93-02, II-G-1,Volume X, Appendix MASS Attachment 16-3 p. 12). The Agency questioned the use of aconstant value for this parameter in view of its potential importance in determining gas and brineflow rates through plugged boreholes and was included it in Enclosure 4 of the Agency’s March19, 1997 letter as requiring further evaluation (Trovato 1997A).

The sensitivity of the PA model to changes in the borehole plug permeability was studied in the

34

Agency’s SA by determining the variation in model performance measures to changes in thevalue of this parameter. The high end of the range (-11 log m2) was determined by the Agencybased on the high end of the range of degraded plug permeabilities estimated by DOE (see CCADocket: A-93-02, II-G-1, Volume X, Appendix MASS Attachment 16-3 p. 14). The low end ofthe range (-18 log m2) is the lowest value measured for the aforementioned WIPP borehole pluggrout in laboratory tests (Christensen and Hunter 1980 p. 4). The results indicated a 101%average change in the performance measures to these changes in the parameter value, which theAgency considered to be significant. The DOE was therefore required by the Agency’s letter ofApril 17, 1997 (Trovato 1997B Enclosure 2) to treat this parameter as a sampled value in thePAVT with a uniform distribution, a lower bound of -19 log m2, and an upper bound of -17 logm2. A uniform distribution for the logarithmic values was selected to provide equal weight acrossthe range of this parameter. The lower bound is one order of magnitude lower than the lowestvalue measured for the WIPP borehole plug grout in the aforementioned laboratory tests. This isconsidered to be a more conservative lower bound because a less permeable borehole plug mayresult in higher repository gas pressures and greater releases during subsequent exploratorydrilling intrusions. The upper bound is equal to the upper bound assumed for concrete in theshaft seals. It is based on the high end of the range of values from laboratory measurements, asadjusted upward to account for uncertainties in field placement techniques and minor concretedeterioration (see WPO # 30640 Section III Vol. 3 Part 2.2.1 p. 1). DOE was also mandated tochange the Y- and Z-direction counterparts of this permeability in the same manner.

5.18 3256 BLOWOUT - FGE: Gravity Scaling Factor

The gravity scaling factor is an experimentally determined parameter used by DOE in theCUTTINGS_S code to relate the limiting fracture erosion velocity to the effects of gravity incalculating spallings release volumes. It is one of three factors applied by DOE to forcesresisting waste erosion during a spallings event. These forces were the gravitational forces on theparticle, capillary forces resulting from partial brine saturation, and the binding forces of particlecementation. The first two of these scaling factors were empirically determined fromexperimental results. The gravity scaling factor was determined to equal 18.1 and the capillaryscaling factor was determined to equal zero (see WPO #35695 p. 32 and Lenke et al. 1996 p. 27).The third factor, the cementation scaling factor, is discussed in Section 3.1 of this report. TheAgency questioned the applicability of the experimental results to spallings releases under WIPPrepository conditions, and considered this parameter to be potentially important because of itsuse in directly calculating a release from the repository. Because of these concerns, thisparameter was included in Enclosure 4 of the Agency’s March 19, 1997 letter as requiringfurther evaluation (Trovato 1997A).

Subsequent to the aforementioned letter, the DOE proposed and the Agency accepted a change inthe way that spallings releases would be addressed in the PAVT that eliminated the need for thisparameter. This change is discussed in Section 3.1 of this report.

5.19 3259 BLOWOUT - APORO: Waste Permeability in CUTTINGS_S Model(663 WAS_AREA - PRMX_LOG)

35

(2131 REPOSIT - PRMX_LOG)

Waste permeability was identified by DOE as the input parameter BLOWOUT - APORO to theCUTTINGS_S model in the model documentation (see WPO #40521 p. 38). This parameter wastreated as a constant in the CCA with a value of 1.7 E-13 m2. The Agency questioned thisparameter because of its potential importance in calculating a direct waste release to the groundsurface. It was therefore included in Enclosure 4 of the Agency’s March 19, 1997 letter asrequiring further evaluation (Trovato 1997A).

DOE was subsequently required by the Agency’s letter of April 25, 1997 (Trovato 1997CEnclosure 2) to change the value of this parameter in the PAVT to 2.4 E-13 m2. This parametervalue is based on a recalculation of the weighted sum of the permeabilities of the individualwaste components using the expected value of a piecewise-linear cumulative distribution ratherthan a uniform distribution (see CCA Docket: A-93-02, II-G-1, Volume XII, Appendix PEER p.5-14). The Agency considers this change to be minor; it was made primarily to correct acomputational error rather than to modify a concept. The Agency did not require this parameterto be treated as a sampled variable based on the reasoning of the DOE’s Engineered SystemsData Qualification Peer Review Panel. The waste permeability, as corrected, was found by thatpanel to be not unreasonable when compared with permeabilities of compacted waste inmunicipal landfills (CCA Docket: A-93-02, II-G-1, Volume XII, Appendix PEER p. 5-18). Also,although that Panel estimated that uncertainties could cause the waste permeability to vary by upto an order of magnitude (CCA Docket: A-93-02, II-G-1, Volume XII, Appendix PEERAttachment PEER-5 p. 5-17), the Agency concluded that because the waste permeability isalready more than two orders of magnitude higher than the permeability of any other geologic orseal component, flow through the waste would be relatively fast and long term releases to theaccessible environment would be fairly insensitive to changes in waste permeability within anorder of magnitude. Although direct brine releases during drilling intrusions would change inproportion to the changes in waste permeability, these releases constitute a small fraction of thetotal radionuclide releases from the repository and changes in their value would not significantlyaffect the mean total release (see CCA Volume I Figure 6-41).Upon further review, the Agency found that despite the identification of this parameter as aninput to CUTTINGS_S, it was not used in the final CCA version of that model. Instead, thewaste permeability was incorporated using parameters 663 (WAS_AREA - PRMX_LOG) and2131 (REPOSIT - PRMX_LOG). Both of these parameters were given constant values of 1.7 E-13 m2 in the CCA. DOE was therefore informed that parameters 663 and 2131 must be changedin the PAVT to 2.4 E-13 m2 in the X-, Y-, and Z-directions (see Kruger 1997). This is the valuelisted for parameter BLOWOUT - APORO in the Agency’s letter of April 25, 1997 (Trovato1997C Enclosure 2).

5.20 3429 PHUMOX3 - PHUMCIM: Humic Colloid Proportionality Constant in CastileBrine

This parameter is the proportionality constant for sorption of +3 actinides to humic colloids inCastile brine. It is defined as the ratio of the moles of actinide sorbed on humic colloids to the

36

moles of actinide remaining in solution. Actinides with an oxidation state of +3 predominate themobile radionuclide inventory and are the most important in evaluating releases. Although onlythe proportionality constant for actinides with an oxidation state of +3 in Castile brine isaddressed here, DOE also assigned proportionality constants for actinides with other oxidationstates and for actinides in Salado brine. The humic colloid proportionality constants weredetermined by DOE largely through literature reviews. These values were assigned to oxidationstates and brine types independent of element type because sorption to large humic moleculesoccurs through chemical complexation and is more sensitive to oxidation state and the chemicalenvironment of the brine than to the specific actinide elements present. Because of its potentialimportance and uncertainty, DOE treated parameter PHUMOX3 - PHUMCIM as a sampledvariable in the CCA with a cumulative distribution, a range of 0.065 to 1.6, and a median of1.37. The Agency considered this parameter to be potentially important in computing brinereleases and included it in Enclosure 4 of the Agency’s March 19, 1997 letter as requiring furtherevaluation (Trovato 1997A).

The sensitivity of the PA model to changes in the humic colloid proportionality constant wasstudied by the Agency in two sensitivity analyses. In the first of these analyses, the values of thisconstant were simultaneously changed along with the values of other parameters associated withactinide solubility (see SA Report 1997 Appendix PD Section PD-4.1). In the second analysis,the values of this constant were varied alone (see SA Report 1997 Appendix PD Section PD-4.3). In both analyses, the humic colloid proportionality constants were changed simultaneouslyfor all actinide oxidation states and for both Castile and Salado brines. The ranges used in thesensitivity analysis were taken from the extremes of the ranges provided in DOE’sdocumentation, and were considered by the Agency to be adequate for testing sensitivity (seeWPO #42248 Table A and SA Report 1997 Appendix PD Sections PD-4.1 and PD-4.3). Theresults of these analyses are summarized in the SA Report (1997 Table 3.4-1). Although anaverage 51% change in the performance measures was found in the first analysis whensolubilities were also changed, only a 3% change was found in the second analysis when onlythe humic colloid proportionality constants were changed. This indicates that the humic colloidproportionality constants were not the primary source of the 51% change in the first analysis, andthat the PA results are not sensitive to changes in these parameters. Because of a low sensitivity,this parameter was therefore identified in the Agency’s letter of April 25, 1997 as being nolonger in question (Trovato 1997C Enclosure 1).

5.21 3471 BLOWOUT - MAXFLOW: Maximum Period of Uncontrolled Borehole Flow

This parameter defines the maximum time period during which uncontrolled brine and gas flowis assumed to occur from an exploration borehole that intrudes the repository. The value of thisparameter was set at 11 days by DOE based on the time that was required to control theproblematic South Culebra Bluff Unit 1 well in January 1978 (see WPO #40520 p. 198 andWPO #43672 p. 10). The effort to control this well is considered by DOE to be a good analog toa hypothetical extreme loss of control during exploratory drilling at the WIPP Site because of itsproximity to WIPP, the similarity of geology, and the time required to mobilize the resources andexpertise to control the well (see Boak et al. 1996 p. A-6). The Agency questioned the value of

37

this parameter because of the small information base from which it was drawn and because of itspotential importance in determining the magnitude of a direct brine release. This parameter wastherefore included in Enclosure 4 of the Agency’s March 19, 1997 letter as requiring furtherevaluation (Trovato 1997A).

The sensitivity of the PA model to changes in the maximum period of uncontrolled boreholeflow was evaluated in the Agency’s SA by determining the change in the performance measurefor high and low values of this parameter. The low value selected by the Agency was 5 days andis equal to the high end of the range for the minimum period of uncontrolled borehole flow (seeSection 5.22 of this report). The high end is equal to 20 days. This value is nearly double theDOE value and was considered by the Agency to provide an adequate test of sensitivity. Theresults showed a change in the performance measure of 197%. Although this sensitivity wasconsidered by the Agency to be significant, upon further evaluation of DOE’s documentation ofdrilling histories in the Delaware Basin, the Agency concluded that DOE’s selection of 11 daysas an upper limit for uncontrolled borehole flow was appropriate because of the close analogiesthat can be made to drilling at the WIPP Site. This parameter was therefore identified in theAgency’s letter of April 25, 1997 as being no longer in question (Trovato 1997C Enclosure 1).

5.22 3472 BLOWOUT - MINFLOW: Minimum Period of Uncontrolled Borehole Flow

This parameter defines the minimum time period during which uncontrolled brine and gas flowis assumed to occur from an exploration borehole that intrudes the repository. The value of thisparameter was set at 3 days by DOE based on the time typically required to drill through theCastile and cement intermediate casing (see WPO #40520 p. 189). The Agency questioned thevalue of this parameter on the same basis as for the maximum period of uncontrolled flow.Lower periods of uncontrolled flow were documented in DOE’s database and this parameter waspotentially important in determining the magnitude of a direct brine release. This parameter wastherefore included in Enclosure 4 of the Agency’s March 19, 1997 letter as requiring furtherevaluation (Trovato 1997A).

The sensitivity of the PA model to changes in the minimum period of uncontrolled borehole flowwas evaluated in the Agency’s SA by determining the change in the performance measure forhigh and low values of this parameter. The low value selected by the Agency was 1 day and thehigh value was 5 days, based on a DOE survey of current drilling practices (see WPO #43672 p.9 and SA Report 1997 Appendix PD Section PD-2.4). The results indicated no sensitivity tochanges in this parameter and this parameter was therefore identified in the Agency’s letter ofApril 25, 1997 as being no longer in question (Trovato 1997C Enclosure 1).

5.23 3433 PHUMOX3 - PHUMSIM: Humic Colloid Proportionality Constant in SaladoBrine

This parameter is the proportionality constant for sorption of +3 actinides to humic colloids inSalado brine. It was defined, determined, and used by DOE in the same way as parameterPHUMOX3 - PHUMCIM for Castile brine discussed in Section 5.20 above; however, parameter

38

PHUMOX3 - PHUMSIM was treated as a constant in the CCA with a value of 0.19 rather thanas a sampled variable. The Agency considered this parameter to be potentially important incomputing brine releases and included it in Enclosure 4 of the Agency’s March 19, 1997 letter asrequiring further evaluation (Trovato 1997A).

The sensitivity of the PA model to changes in this humic colloid proportionality constant wasstudied by the Agency in the same two sensitivity analyses described in Section 5.20 and in theSA Report (1997). The results indicated that the PA results are not sensitive to changes in thehumic colloid proportionality constants. Because of a low sensitivity, this parameter wasidentified in the Agency’s letter of April 25, 1997 as being no longer in question (Trovato 1997CEnclosure 1).

5.24 3470 BLOWOUT - GAS_MIN: DBR Cutoff Gas Flow Rate

Direct brine releases are assumed by DOE to occur over a minimum of 3 days and then ceasewhen either the gas flow rate drops below the value prescribed by this cutoff gas flow rateparameter, or when flow occurs in excess of 11 days, whichever occurs first. The cutoff gas flowrate was treated by DOE as a constant with a value of 100 mscf/day (100,000 standard cubic feetper day), which DOE considered to be a rate that a driller could readily control (see WPO#40520 p. 189). The Agency questioned the value of this parameter because of its lack of adocumented source and because of its potential importance in determining the magnitude of adirect brine release. This parameter was therefore included in Enclosure 4 of the Agency’sMarch 19, 1997 letter as requiring further evaluation (Trovato 1997A).

The sensitivity of the PA model to changes in the cutoff gas flow rate was evaluated in theAgency’s SA by determining the change in the performance measure for high and low values ofthis parameter. The low value selected by the Agency was 50 mscf/day and is equal to half of theCCA value. The high end was 200 mscf/day and is twice the CCA value (see SA Report 1997Appendix PD Section PD-2.2). This range was considered by the Agency to provide an adequatetest of sensitivity. The results showed no sensitivity to changes in the value of this parameter(see SA Report 1997 Table 3.2-1). This parameter was therefore identified in the Agency’s letterof April 25, 1997 as being no longer in question (Trovato 1997C Enclosure 1).

39

5.25 3317 PU - PROPMIC: Microbial Colloid Proportionality Constant for Plutonium

This parameter is the proportionality constant for sorption of plutonium ions to microbialcolloids. It is defined as the ratio of the moles of plutonium sorbed on microbial colloids to themoles of plutonium remaining in solution. Plutonium and americium dominate the mobileradionuclide inventory and are the most important in evaluating releases. Values of thisproportionality constant for americium are discussed in Section 5.33 below. Values for thisproportionality constant were determined based on experiments conducted for the WIPP atBrookhaven National Laboratory. These values are assigned to actinides independent ofoxidation state and brine type because, for living organisms, actinide uptake and toxicity aremore closely associated with element type than with these variables. The parameter PU -PROPMIC was treated as a constant in the CCA with a value of 0.3, although a range of valuesfor this parameter was identified in DOE’s documentation (see WPO #35856 Table 1). TheAgency considered this parameter to be potentially important in computing actinide solubilityand questioned whether its uncertainty had been adequately captured in the CCA. This parameterwas therefore included in Enclosure 4 of the Agency’s March 19, 1997 letter as requiring furtherevaluation (Trovato 1997A).

The sensitivity of the PA model to changes in the microbial colloid proportionality constants wasstudied by the Agency in two sensitivity analyses. In the first of these analyses, the values ofthis constant were simultaneously changed along with the values of other parameters associatedwith actinide solubility (see SA Report 1997 Appendix PD Section PD-4.1). In the secondanalysis, the values of this constant were varied alone (see SA Report 1997 Appendix PDSection PD-4.3). In both analyses the proportionality constants for both plutonium andamericium were simultaneously assigned high or low values equal to the extremes of the rangesidentified in the aforementioned DOE documentation. The results of these analyses aresummarized in the SA Report (1997 Table 3.4-1). Although an average 51% change in theperformance measures was found in the first analysis when other parameters were also changed,only a 4% change was found in the second analysis when only the microbial colloidproportionality constants were changed. This indicates that the microbial colloid proportionalityconstants were not the primary source of the 51% change in the first analysis, and that the PAresults are not sensitive to changes in these parameters. Because of a low sensitivity, thisparameter was therefore identified in the Agency’s letter of April 25, 1997 as being no longer inquestion (Trovato 1997C Enclosure 1).

5.26 3405 SOLMOD6 - SOLCIM: U (VI) Solubility in Castile Brine

This parameter is the expected solubility of uranium in the +6 oxidation state in Castile brine.The same constant value of 8.8 E-6 moles/liter, based on literature reviews, was used by DOE inthe CCA for both Castile and Salado brines. The Agency considered this parameter to bepotentially important in computing the uranium content of brine releases and questioned whetherits value was applicable to WIPP brines. This parameter was therefore included in Enclosure 4 ofthe Agency’s March 19, 1997 letter as requiring further evaluation (Trovato 1997A).

40

The sensitivity of the sampled solubility calculated in the PA model to changes in the principalinput parameters was studied by the Agency by simultaneously varying the values of thoseparameters for all actinides, oxidation states, and brines (see SA Report 1997 Appendix PDSection PD-4.1). For uranium +6 these were the SOLU6 - SOLSIM and SOLU6 - SOLCIMparameters (defined below), the microbial colloid proportionality constants (PROPMIC; seeSections 5.25 and 5.33), and the humic colloid proportionality constants (PHUMSIM andPHUMCIM; see Sections 5.20 and 5.23). SOLU6 - SOLSIM and SOLU6 - SOLCIM are definedin the following equations:

For uranium +6 in Salado brine:

Sampled Solubility used in CCA = SOLMOD6-SOLSIM x 10SOLU6-SOLSIM

For uranium +6 in Castile brine:

Sampled Solubility used in CCA = SOLMOD6-SOLCIM x 10SOLU6-SOLCIM

SOLU6 - SOLSIM and SOLU6 - SOLCIM are therefore high and low values used as powers often to define the range over which uranium +6 solubility was sampled by DOE for the CCA. Theparameters SOLMOD6 - SOLSIM and SOLMOD6 - SOLCIM are the expected uranium +6solubility values for Salado and Castile brines, respectively. As shown in the equations,solubility values are sampled by combining parameters for both range of solubility and expectedsolubility value, rather than by sampling directly from a single parameter. The same range ofvalues was used by DOE for both SOLU6 - SOLSIM and SOLU6 - SOLCIM as well as for allother actinides, oxidation states, and brines, reflecting a lack of detailed knowledge of theseparameters.

Because of the potential importance of parameter SOLMOD6 - SOLCIM in calculating directradionuclide releases and the Agency’s questions regarding the technical approach taken byDOE to determine the expected solubility of uranium +6 in Castile brine, a mandated constantvalue of 4.6 E-3 moles/liter was identified for this parameter in the Agency’s letter of April 25,1997 for use in the PAVT (Trovato 1997C Enclosure 2). The Agency did not provide a range ofvalues for this parameter because sampling of uranium +6 solubilities is conducted by samplinganother parameter, SOLU6 - SOLCIM, as described above. For similar reasons, the Agency alsodid not provide a range of values for the other SOLMOD parameters discussed below. TheAgency’s mandated value was calculated for Castile brine based on the solubility of the uraniummineral schoepite (UO3

.2H2O) using the USGS PHREEQC model and assuming chemicalconditions based on equilibrium between brucite and hydromagnesite (EPA, 1998c, Docket No.A-93-02, V-B-17). This calculation provided a conservative estimate because the uranium +6concentrations calculated for schoepite solubility are greater than the concentrations measured inDOE’s brine experiments. These brine experiments were conducted under alkaline conditionsexpected to be representative of the repository environment. In view of the lack of other WIPP-specific values, the Agency considered this approach acceptable. The Agency did not mandatechanges in SOLU6 - SOLCIM for the PAVT because the range of values for that sampled

41

parameter was considered sufficiently broad to capture the uncertainty in parameter SOLMOD6- SOLCIM.

Subsequent to the Agency’s aforementioned April 25, 1997 letter, additional experimental results provided by DOE indicated that uranium +6 solubilities in high ionic strength WIPPbrines were lower than the value of 8.8 E-6 moles/liter used in the CCA (see Kruger 1997; WPO# 44625; WPO # 45115). Based on this new information as well as the original basis used byDOE to establish the CCA parameter values (see WPO # 36488), the Agency informed DOE thatthe uranium +6 solubilities used in the CCA for Castile and Salado brines (parameters 3405 and3409) were considered adequate and did not need to be changed in the PAVT.

5.27 3409 SOLMOD6 - SOLSIM: U(VI) Solubility in Salado Brine

This parameter is the expected solubility of uranium in the +6 oxidation state in Salado brine.The same constant value of 8.8 E-6 moles/liter, based on literature reviews, was used by DOE inthe CCA for both Castile and Salado brines. The Agency considered this parameter as well as theparallel parameter for Castile brine (see Section 5.26) to be potentially important in computingthe uranium content of brine releases and questioned whether the assigned value was applicableto WIPP brines. This parameter was therefore included in Enclosure 4 of the Agency’s March19, 1997 letter as requiring further evaluation (Trovato 1997A).

Because of the potential importance of the parameter SOLMOD6 - SOLSIM in calculating directradionuclide releases and the Agency’s questions regarding the technical approach taken byDOE to determine the expected solubility of uranium +6 in Salado brine, a mandated value of3.7 E-5 moles/liter was identified for this parameter in the Agency’s letter of April 25, 1997 foruse in the PAVT (Trovato 1997C Enclosure 2). The Agency’s mandated value was calculated forSalado brine in the same manner as described for Castile brine in Section 5.26. This calculationprovided a conservative estimate as mentioned above. However, as described in Section 5.26,subsequent to the Agency’s aforementioned April 25, 1997 letter, additional experimental resultsprovided by DOE allowed the Agency to conclude that the uranium +6 solubilities used in theCCA for Castile and Salado brines (parameters 3405 and 3409) were adequate and did not needto be changed in the PAVT.

5.28 3402 SOLMOD3 - SOLCIM: Oxidation State +III Solubility in Castile Brine(3406 SOLMOD3 - SOLSIM)

This parameter is used by DOE in the CCA as the expected solubility of all actinides in the +3oxidation state in Castile brine. The solubility of americium +3 in Castile brine was used byDOE as a surrogate for the solubility of other +3 actinides because of their limited or inadequatedata base. A constant value of 6.52 E-8 moles/liter, based on FMT computer code modeling, wasused by DOE for this parameter. The Agency’s analysis of the FMT code used to calculatesolubilities revealed incorrect solubility estimates because the database used as an input to FMTwas in error, as were the resulting CCA calculations (EPA 1998d, Appendix A). Because ofthese errors and because of the potential importance of this parameter in calculating the

42

concentration of +3 actinides in brine releases, it was included in Enclosure 4 of the Agency’sMarch 19, 1997 letter as requiring further evaluation (Trovato 1997A).

DOE assumed in the CCA that, under the expected reducing repository conditions, plutoniumdissolved in the Castile and Salado brines would exist predominantly in either the +3 or +4oxidation states and assigned equal probabilities to each of these states in the PA. The repositoryinventory includes large quantities of reducing agents in the form of metallic iron, organicmatter, and organic chemicals. These reductants are expected to rapidly consume any availableoxygen present in the repository atmosphere shortly after closure, producing reducingconditions. Under reducing conditions, plutonium is expected to be stable in either the +3 or +4oxidation states based on chemical equilibrium arguments. In confirmation of this expectation,experimental studies conducted by DOE have shown that Pu +6 is rapidly reduced to loweroxidation states by direct reactions with iron and organic compounds (WPO # 35197). Also,stabilization of plutonium in the +6 oxidation state by carbonate is not expected to occur in therepository because reactions with the magnesium oxide backfill should limit CO2 (g) to very lowpartial pressures based on chemical equilibrium calculations.

Experimental studies indicate that both Pu +3 and Pu +4 are produced by the reduction of Pu +6with various metallic iron and organic compounds. However, the predominant oxidation stateexpected for the repository conditions cannot be defined with absolute certainty based on theexperimental studies or equilibrium model calculations. Consequently, DOE assumed that bothPu +3 and Pu +4 were possible and included both in the PA calculations.

Upon review, the Agency concurs with DOE’s conclusions regarding plutonium oxidation statesand believes that the foregoing approach is reasonable and consistent with current knowledge ofplutonium redox chemistry and expected repository conditions.

A mandated value of 1.38 E-8 moles/liter was identified for parameter SOLMOD3 - SOLCIM inthe Agency’s letter of April 25, 1997 for use in the PAVT (Trovato 1997C Enclosure 2). Thismandated value was calculated for Castile brine using the corrected database to execute the FMTmodel and assuming chemical conditions based on equilibrium between brucite andhydromagnesite rather than between brucite and magnesite (EPA 1998d, Appendix A, Table 4).This calculation is based on the Pitzer activity coefficient formalism and is consideredappropriate for high ionic strength solutions. As a result of the model corrections, the calculatedsolubility was lower than that used by DOE in the CCA. This change will result in smallerconcentrations of actinides in solution and decreased actinide releases.

Parameter 3406 SOLMOD3 - SOLSIM is the expected solubility of all actinides in the +3oxidation state in Salado brine. Although it was inadvertently not included in the Agency’sMarch 19, 1997 letter as requiring further evaluation (Trovato 1997A), DOE was required by theAgency’s letter of April 25, 1997 to use an alternate value for this parameter in the PAVT(Trovato 1997C Enclosure 2). The solubility of americium +3 in Salado brine was used by DOEas a surrogate for the solubility of other +3 actinides because of their limited or inadequate database. A constant value of 5.82 E-8 moles/liter, based on FMT modeling, was used by DOE for

43

this parameter. The Agency’s concerns for this parameter stem from the aforementioneddatabase errors used to execute the FMT model (EPA 1998d, Appendix A) and the potentialimportance of this parameter in calculating the actinide content of brine releases. A mandatedvalue of 1.2 E-7 moles/liter was identified for parameter SOLMOD3 - SOLSIM in the Agency’sletter of April 25, 1997 for use in the PAVT (Trovato 1997C Enclosure 2). The Agency’smandated value was calculated for Salado brine using the corrected database used to execute theFMT model and assuming chemical conditions based on equilibrium between brucite andhydromagnesite (EPA 1998d, Appendix A, Table 3). This calculation is based on the Pitzeractivity coefficient formalism and is considered appropriate for high ionic strength solutions. Asa result of the model corrections, the calculated solubility was higher than that used by DOE inthe CCA. This change will result in greater concentrations of actinides in solution and increasedactinide releases. Ranges for the two SOLMOD3 parameters were not specified by the Agencybecause solubility is sampled in the CCA using other parameters, as explained in Section 5.26.

5.29 3403 SOLMOD4 - SOLCIM: Oxidation State +IV Solubility in Castile Brine

This parameter is used by DOE in the CCA as the expected solubility of all actinides in the +4oxidation state in Castile brine. The solubility of thorium +4 was used by DOE as a surrogate forthe solubility of other +4 actinides because of their limited or inadequate data base. A constantvalue of 6.00 E-9 moles/liter, based on FMT modeling, was used by DOE for this parameter. Asmentioned in Section 5.28, the Agency’s analysis of the FMT computer code used to calculatesolubilities revealed incorrect solubility estimates because the database used for as an input toFMT was in error, as were the resulting CCA calculations (EPA 1998d, Appendix A). Becauseof these errors and because of the potential importance of this parameter in calculating theconcentration of +4 actinides in brine releases, it was included in Enclosure 4 of the Agency’sMarch 19, 1997 letter as requiring further evaluation (Trovato 1997A).

A mandated value of 4.1 E-8 moles/liter was identified for parameter SOLMOD4 - SOLCIM inthe Agency’s letter of April 25, 1997 for use in the PAVT (Trovato 1997C Enclosure 2). TheAgency’s mandated value was calculated for Castile brine using the corrected FMT databasevalues and assuming chemical conditions based on equilibrium between brucite andhydromagnesite (EPA 1998d, Appendix A, Table 4). This calculation is based on the Pitzeractivity coefficient formalism and is considered appropriate for high ionic strength solutions. Asa result of the model corrections, the calculated solubility was higher than that used by DOE inthe CCA. This change will result in greater concentrations of actinides in solution and increasedactinide releases. A range for this SOLMOD4 parameter was not specified by the Agencybecause solubility is sampled in the CCA using other parameters, as explained in Section 5.26.

5.30 3407 SOLMOD4 - SOLSIM: Oxidation State +IV Solubility in Salado Brine

This parameter is used by DOE in the CCA as the expected solubility of all actinides in the +4oxidation state in Salado brine. The solubility of thorium +4 was used by DOE as a surrogate forthe solubility of other +4 actinides because of their limited or inadequate data base. A constantvalue of 4.4 E-6 moles/liter, based on FMT modeling, was used by DOE for this parameter. As

44

mentioned in Section 5.28, the Agency’s analysis of the FMT code as used to calculatesolubilities revealed incorrect solubility estimates because the database used for as an input toFMT was in error, as were the resulting CCA calculations (EPA 1998d, Appendix A). Becauseof these errors and because of the potential importance of this parameter in calculating theconcentration of +4 actinides in brine releases, it was included in Enclosure 4 of the Agency’sMarch 19, 1997 letter as requiring further evaluation (Trovato 1997A).

Because of the potential importance of this parameter in calculating radionuclide releases and toprovide an opportunity to use the corrected database used to execute the FMT model inperformance assessment, a mandated value of 1.3 E-8 moles/liter was identified for parameterSOLMOD4 - SOLSIM in the Agency’s letter of April 25, 1997 for use in the PAVT (Trovato1997C Enclosure 2). The Agency’s mandated value was calculated for Salado brine using thecorrected FMT model and assuming chemical conditions based on equilibrium between bruciteand hydromagnesite (EPA 1998d, Appendix A, Table 3). This calculation is based on the Pitzeractivity coefficient formalism and is considered appropriate for high ionic strength solutions. Asa result of the model corrections, the calculated solubility was lower than that used by DOE inthe CCA. This change will result in smaller concentrations of actinides in solution and decreasedactinide releases. A range for this SOLMOD4 parameter was not specified by the Agencybecause solubility is sampled in the CCA using other parameters, as explained in Section 5.26.

5.31 3404 SOLMOD5 - SOLCIM: Oxidation State +V Solubility in Castile Brine

This parameter is used by DOE in the CCA as the expected solubility of all actinides in the +5oxidation state in Castile brine. The solubility of neptunium +5 was used by DOE as a surrogatefor the solubility of other +5 actinides because of their limited or inadequate data base. Aconstant value of 2.2 E-6 moles/liter, based on FMT modeling, was used by DOE for thisparameter. As mentioned in Section 5.28, the Agency’s analysis of the FMT code used tocalculate solubilities revealed incorrect solubility estimates because the database used as an inputto FMT was in error, as were the resulting CCA calculations (EPA 1998d, Appendix A).Because of these errors and because of the potential importance of this parameter in calculatingthe concentration of +5 actinides in brine releases, it was included in Enclosure 4 of theAgency’s March 19, 1997 letter as requiring further evaluation (Trovato 1997A).

A mandated value of 4.8 E-7 moles/liter was identified for parameter SOLMOD5 - SOLCIM inthe Agency’s letter of April 25, 1997 for use in the PAVT (Trovato 1997C Enclosure 2). TheAgency’s mandated value was calculated for Castile brine using the corrected FMT model andassuming chemical conditions based on equilibrium between brucite and hydromagnesite (EPA1998d, Appendix A, Table 4). This calculation is based on the Pitzer activity coefficientformalism and is considered appropriate for high ionic strength solutions. As a result of themodel corrections, the calculated solubility was lower than that used by DOE in the CCA. Thischange will result in smaller concentrations of actinides in solution and decreased actinidereleases. A range for this SOLMOD5 parameter was not specified by the Agency becausesolubility is sampled in the CCA using other parameters, as explained in Section 5.26.

45

5.32 3408 SOLMOD5 - SOLSIM: Oxidation State +V Solubility in Salado Brine

This parameter is used by DOE in the CCA as the expected solubility of all actinides in the +5oxidation state in Salado brine. The solubility of neptunium +5 was used by DOE as a surrogatefor the solubility of other +5 actinides because of their limited or inadequate data base. Aconstant value of 2.3 E-6 moles/liter, based on FMT modeling, was used by DOE for thisparameter. As mentioned in Section 5.28, the Agency’s analysis of the FMT code to calculatesolubilities revealed incorrect solubility estimates because the database used for as an input toFMT was in error, as were the resulting CCA calculations (EPA 1998d, Appendix A). Becauseof these errors and because of the potential importance of this parameter in calculating theconcentration of +5 actinides in brine releases, it was included in Enclosure 4 of the Agency’sMarch 19, 1997 letter as requiring further evaluation (Trovato 1997A).

A mandated value of 2.4 E-7 moles/liter was identified for parameter SOLMOD4 - SOLSIM inthe Agency’s letter of April 25, 1997 for use in the PAVT (Trovato 1997C Enclosure 2). TheAgency’s mandated value was calculated for Salado brine using the corrected FMT model andassuming chemical conditions based on equilibrium between brucite and hydromagnesite (EPA1998d, Appendix A, Table 3). This calculation is based on the Pitzer activity coefficientformalism and is considered appropriate for high ionic strength solutions. As a result of themodel corrections, the calculated solubility was lower than that used by DOE in the CCA. Thischange will result in smaller concentrations of actinides in solution and decreased actinidereleases. A range for this SOLMOD5 parameter was not specified by the Agency becausesolubility is sampled in the CCA using other parameters, as explained in Section 5.26.

5.33 3311 AM - PROPMIC: Microbial Colloid Proportionality Constant for Americium

This parameter is the proportionality constant for sorption of americium ions to microbialcolloids. It is defined as the ratio of the moles of americium sorbed on microbial colloids to themoles of americium remaining in solution. This parameter is treated in the same manner asparameter PU - PROPMIC described in Section 5.25 above. Plutonium and americium dominatethe mobile radionuclide inventory and are the most important in evaluating releases. Theparameter AM - PROPMIC was treated as a constant in the CCA with a value of 3.6, although arange of values for this parameter was identified in DOE’s documentation (see WPO #35856Table 1). The Agency considered this parameter to be potentially important in computingactinide solubility and questioned whether its uncertainty had been adequately captured in theCCA. This parameter was then included in Enclosure 4 of the Agency’s March 19, 1997 letter asrequiring further evaluation (Trovato 1997A).

The sensitivity of the PA model to changes in the microbial colloid proportionality constants wasstudied by the Agency in two sensitivity analyses described in Section 5.25. In both analyses theproportionality constants for both plutonium and americium were simultaneously assigned highor low values corresponding to the extremes of the ranges identified in the aforementioned DOEdocumentation. The results of these analyses are summarized in the SA Report (1997 Table 3.4-1). Although an average 51% change in the performance measures was found in the first analysis

46

when other parameters were also changed, only a 4% change was found in the second analysiswhen only the microbial colloid proportionality constants were changed. This indicates that themicrobial colloid proportionality constants were not the primary source of the 51% change in thefirst analysis, and that the PA results are not sensitive to changes in these parameters. Because ofa low sensitivity, this parameter was therefore identified in the Agency’s letter of April 25, 1997as being no longer in question (Trovato 1997C Enclosure 1).

5.34 3482 AM+3 - MKD_AM: Matrix Partition Coefficient for Americium +III

This parameter is the solid-liquid partition coefficient (Kd) for americium +3 in the Culebradolomite. DOE developed values for this parameter based on laboratory experiments performedat Los Alamos National Laboratory and Sandia National Laboratories (see WPO #38801). Thisparameter was treated as a sampled variable in the CCA with a uniform distribution and a rangeof 0.02 m3/kg to 0.5 m3/kg. Because of the potential importance of this and other Kd values inpredicting the migration of actinides in the Culebra dolomite to the accessible environment, adetailed evaluation of DOE’s methodology for determining Kd values was conducted by theAgency (EPA, 1997b). Concerns identified in the Agency’s evaluation, discussed below, causedthe Agency to question DOE’s technical approach, and because of the importance of americiumin WIPP’s actinide inventory, included the Kd for americium +3 in Enclosure 4 of the Agency’sMarch 19, 1997 letter as requiring further evaluation (Trovato 1997A).

In the aforementioned Agency evaluation (EPA, 1997b), the actinide Kds used by DOE in theCCA were compared with literature values, DOE’s experimental approach was criticallyreviewed, and DOE’s experiments on the effects of organic ligands were reviewed. The literaturereview indicated that in most cases the actinide Kds determined for the WIPP brines fell into thelower end of the range of values reported for sands, loams, and clayey soils. However, lessretention of the actinides on solids in a high ionic strength brine would be expected than in moredilute solutions because of swamping of available adsorption sites by the high concentrations ofthe ionic species present in the brine, and also because of the lower sorptive capacity of thedolomitic rock (see EPA, 1997b). Lower Kd values are more conservative because higheractinide concentrations are predicted to remain in solution and remain available for transport.

In reviewing DOE’s experimental work on Kds, the Agency concluded that the experimentalprocedure, including the use of crushed Norwegian dolomite as well as crushed Culebradolomite, was adequate. Crushed samples are often used in Kd experiments to minimizeimprecision resulting from variabilities in surface chemical properties, and given the long timeperiod over which actinide migration occurs in the Culebra, DOE’s assertion that fluid movingthrough this dolomite will eventually be exposed to the same surface to volume ratio that wasused in the experiments is not unreasonable. Also, the Agency agrees with DOE’s objective ofusing the purer Norwegian dolomite to provide more conservative measures of Kds by focusingon the mineral dolomite and reducing the effects of impurities such as clays in the Culebradolomite that could tend to increase the measured values.

However, the Agency identified a lack of documentation that equilibrium had been reached in

47

the experiments and that the effects of actinide concentrations on the measured Kds wereconsidered (see EPA, 1997b). Although this lack of documentation was of concern, it was notconsidered a major technical issue. The DOE’s experiments were continued for 8 weeks ratherthan the 48 hours or less normally used for equilibration in Kd experiments, and are consideredby the Agency to have an adequate duration. Also, the low actinide concentrations used in theexperiments essentially precluded the formation of precipitates and the loss of actinides fromsolution by that mechanism.

The Agency also noted that the effects of the MgO backfill and the more alkaline brine solutionit would create were not considered in the experiments. Although there is some uncertainty aboutthe effects of a more alkaline solution on adsorption to dolomite, for which less information isavailable than for other geologic media, the adsorption of cationic species generally increaseswith increasing pH. Calcite, which is chemically similar to dolomite and may be considered ananalog for that mineral, does exhibit this tendency for increasing adsorption with increasing pH(see EPA, 1997b). The use of a lower pH brine in the Kd experiments than would actually bepresent in the repository is therefore expected to underestimate Kd values and provide moreconservative calculations of potential releases by increasing the radionuclide fraction remainingdissolved in the brine.

The uncertainty of DOE’s experimental results was increased when the Kd experiments foramericium +3 failed (for the reasons described in WPO # 38801), and Kds measured forplutonium +5 were used to represent americium +3. Also, because americium +3 was to be theanalog for plutonium +3, the Kds measured for plutonium +5 were also used for plutonium +3.Although the Agency would prefer to not use surrogates for these parameters, particularlysurrogates with different oxidation states, the uncertainty in these parameters has beenadequately captured within the aforementioned range of 0.02 to 0.5 m3/kg sampled in PA. The Agency’s review of DOE’s experimental work on organic ligands indicated that although atlow ligand concentrations some actinide Kds increased relative to values determined in theabsence of organics, at higher ligand concentrations actinide Kds generally showed decreasedvalues. The results of the ligand experiments were not used by DOE in deriving the Kd rangesused in the CCA based on the hypothesis that a surplus of metallic cations in the repository brinewill form complexes with the ligands, making them generally unavailable for affecting actinidespeciation. DOE tested the validity of this hypothesis by using the FMT model to performspeciation calculations comparing the effects of organic ligands in the experimental solutionswith those expected under repository conditions (Novak et al. 1996). Results from the FMTcalculations were found to be generally consistent with the aforementioned hypothesis.

Although the Agency recognizes that the FMT calculations are not completely descriptive of theexpected repository conditions, consideration of Kds reported in the literature, the expectedincreased adsorption under alkaline conditions resulting from MgO backfill reactions, and theresults of the speciation calculations support use of the Kd ranges used by DOE in the CCA (seeEPA, 1997b). However, the Kd values obtained by DOE appear to be logarithmically distributed,and the Agency questions DOE’s use of a uniform distribution for Kds in the CCA (EPA, 1997b).Because of the Agency’s question regarding the form of the distribution, and because the

48

actinide Kds range over more than an order of magnitude, DOE was required by the Agency’sletter of April 25, 1997 (Trovato 1997C Enclosure 2) to treat this parameter as a sampled valuein the PAVT with a loguniform rather than a uniform distribution, while maintaining the samerange of values as used in the CCA. Use of a loguniform distribution provides equal weight insampling across the range of values and conservatively reduces the median value of thisparameter, thereby increasing the average mass of actinides in solution.

5.35 3480 PU+3 - MKD_PU: Matrix Partition Coefficient for Plutonium +III

This parameter is the solid-liquid partition coefficient (Kd) for plutonium +3 in the Culebradolomite. DOE developed values for this parameter based on laboratory experiments usingplutonium +5 as an analog (see WPO #38801). This parameter was treated as a sampled variablein the CCA with a uniform distribution and a range of 0.02 m3/kg to 0.5 m3/kg. This is the samerange as for americium +3, for which plutonium +5 was also an analog. A detailed discussion ofthe Agency’s review and conclusions regarding DOE’s Kd values is presented in Section 5.34above. Because of the potential importance of this and other Kd values in predicting themigration of actinides in the Culebra dolomite to the accessible environment, and because ofconcerns identified in the Agency’s evaluation of the technical basis for this parameter, the Kdfor plutonium +3 was included in Enclosure 4 of the Agency’s March 19, 1997 letter as requiringfurther evaluation (Trovato 1997A).

An overview of the Agency’s technical evaluation of DOE’s Kd values is presented in Section5.34. As a result of this evaluation, the Agency concluded that the ranges of Kd values used inthe CCA were appropriate and adequately reflected the uncertainty in this parameter. However,the Agency questioned DOE’s use of a uniform distribution for Kds in the CCA. Previousreviews of Kds reported in the literature have found this parameter to be logarithmicallydistributed (see EPA, 1997b), and because the actinide Kds range over more than an order ofmagnitude, DOE was required by the Agency’s letter of April 25, 1997 (Trovato 1997CEnclosure 2) to treat this parameter as a sampled value in the PAVT with a loguniform ratherthan a uniform distribution, while maintaining the same range of values as used in the CCA. Useof a loguniform distribution provides equal weight in sampling across the range of values andconservatively reduces the median value of this parameter, thereby increasing the average massof actinides in solution.

5.36 3481 PU+4 - MKD_PU: Matrix Partition Coefficient for Plutonium +IV

This parameter is the solid-liquid partition coefficient (Kd) for plutonium +4 in the Culebradolomite. DOE developed values for this parameter based on laboratory experiments usingthorium +4 as an analog (see WPO #38801). This parameter was treated as a sampled variable inthe CCA with a uniform distribution and a range of 0.9 m3/kg to 20 m3/kg. This is the same rangeas for uranium +4, for which thorium +4 (Parameter 3478 TH+4 - MKD_TH) was also ananalog. A detailed discussion of the Agency’s review and conclusions regarding DOE’s Kdvalues is presented in Section 5.34 above. Because of the potential importance of this and otherKd values in predicting the migration of actinides in the Culebra dolomite to the accessible

49

environment, and because of concerns identified in the Agency’s evaluation of the technicalbasis for this parameter, the Kd for plutonium +4 was included in Enclosure 4 of the Agency’sMarch 19, 1997 letter as requiring further evaluation (Trovato 1997A).


5.37 3479 U+4 - MKD_U: Matrix Partition Coefficient for Uranium +IV

This parameter is the solid-liquid partition coefficient (Kd) for uranium +4 in the Culebradolomite. DOE developed values for this parameter based on laboratory experiments usingthorium +4 as an analog (see WPO #38801). This parameter was treated as a sampled variable inthe CCA with a uniform distribution and a range of 0.9 m3/kg to 20 m3/kg. This is the same rangeas for plutonium +4, for which thorium +4 was also an analog. A detailed discussion of theAgency’s review and conclusions regarding DOE’s Kd values is presented in Section 5.34 above.Because of the potential importance of this and other Kd values in predicting the migration ofactinides in the Culebra dolomite to the accessible environment, and because of concernsidentified in the Agency’s evaluation of the technical basis for this parameter, the Kd for uranium+4 was included in Enclosure 4 of the Agency’s March 19, 1997 letter as requiring furtherevaluation (Trovato 1997A).


50

5.38 3475 U+6 - MKD_U: Matrix Partition Coefficient for Uranium +VI

This parameter is the solid-liquid partition coefficient (Kd) for uranium +6 in the Culebradolomite. DOE developed values for this parameter based on laboratory experiments (see WPO#38801). This parameter was treated as a sampled variable in the CCA with a uniformdistribution and a range of 0.00003 m3/kg to 0.03 m3/kg. A detailed discussion of the Agency’sreview and conclusions regarding DOE’s Kd values is presented in Section 5.34. Because of thepotential importance of this and other Kd values in predicting the migration of actinides in theCulebra dolomite to the accessible environment, and because of concerns identified in theAgency’s evaluation of the technical basis for this parameter, the Kd for uranium +6 wasincluded in Enclosure 4 of the Agency’s March 19, 1997 letter as requiring further evaluation(Trovato 1997A).

An overview of the Agency’s technical evaluation of DOE’s Kd values is presented in Section5.34. As a result of this evaluation, the Agency concluded that the ranges of Kd values used inthe CCA were appropriate and adequately reflected the uncertainty in this parameter. However,the Agency questioned DOE’s use of a uniform distribution for Kds in the CCA. Previousreviews of Kds reported in the literature have found this parameter to be logarithmicallydistributed (see EPA, 1997b), and because the actinide Kds range over more than an order ofmagnitude, DOE was required by the Agency’s letter of April 17, 1997 (Trovato 1997BEnclosure 2) to treat this parameter as a sampled value in the PAVT with a loguniform ratherthan a uniform distribution, while maintaining the same range of values as used in the CCA. Useof a loguniform distribution provides equal weight in sampling across the range of values andconservatively reduces the median value of this parameter, thereby increasing the average massof actinides in solution.

5.39 656 WAS_AREA - GRATMICH: Gas Generation Rate due to Microbial Actionunder Humid Conditions

The gas generation due to microbial action under humid (brine unsaturated) conditions wasestimated by DOE based on laboratory studies of microbial consumption of cellulosics atBrookhaven National Laboratory (see CCA Docket: A-93-02, II-G-1, Volume X, AppendixMASS p. MASS-55). DOE treated this parameter as a sampled variable in the CCA with auniform distribution and a range of zero to 1.268 E-9 moles/kg-s (see WPO #34923 p. 1).Because of the potential importance of this parameter in calculating repository gas pressures, theAgency questioned whether its uncertainty had been adequately captured in the CCA. Thisparameter was therefore included in Enclosure 4 of the Agency’s March 19, 1997 letter asrequiring further evaluation (Trovato 1997A).

The sensitivity of the PA model to changes in the gas generation due to microbial action underhumid conditions was studied by the Agency in a sensitivity analysis (see SA Report 1997Appendix PD Section PD-1.22). In this analysis the value of this parameter was set equal to theextreme values of the range sampled by DOE in the CCA. No sensitivity of the performancemeasures to changes in this parameter were found (see SA Report 1997 Table 3.1-1). Because of

51

a lack of sensitivity, this parameter was identified in the Agency’s letter of April 25, 1997 asbeing no longer in question (Trovato 1997C Enclosure 1).

5.40 657 WAS_AREA - GRATMICI: Gas Generation Rate due to Microbial Actionunder Inundated Conditions

The gas generation due to microbial action under inundated (brine saturated) conditions wasestimated by DOE based on laboratory studies of microbial consumption of cellulosics atBrookhaven National Laboratory (see CCA Docket: A-93-02, II-G-1, Volume X, AppendixMASS p. MASS-55). DOE treated this parameter as a sampled variable in the CCA with auniform distribution and a range of 3.17 E-10 mole/kg-s to 9.51 E-9 moles/kg-s (see CCADocket: A-93-02, II-G-1, Volume XI, Appendix PAR p. PAR-21 and WPO #34928 p. 1).Because of the potential importance of this parameter in calculating repository gas pressures, theAgency questioned whether its uncertainty had been adequately captured in the CCA. Thisparameter was therefore included in Enclosure 4 of the Agency’s March 19, 1997 letter asrequiring further evaluation (Trovato 1997A).

The sensitivity of the PA model to changes in the gas generation due to microbial action underinundated conditions was studied by the Agency in a sensitivity analysis (see SA Report 1997Appendix PD Section PD-1.23). In this analysis the value of this parameter was set equal to theextreme values of the range sampled by DOE in the CCA. The average change in theperformance measures resulting from changes in the value of this parameter was found to be 1%and was not considered significant by the Agency (see SA Report 1997 Table 3.1-1). Because ofthis low sensitivity, this parameter was identified in the Agency’s letter of April 17, 1997 asbeing no longer in question (Trovato 1997B Enclosure 1).

52

6.0 SUMMARY AND CONCLUSIONS

This report describes the disposition of 58 parameters used by DOE in the CCA PA that werefound to be inadequately supported following an extensive review of DOE’s parameter databaseby the Agency. These parameters were identified to DOE in Enclosures 2, 3, and 4 of theAgency’s letter of March 19, 1997 and are discussed in Sections 3, 4, and 5, respectively, of thisreport (Trovato 1997A). These parameters were identified because they were potentiallyimportant to the results of the PA and because they lacked supporting data, they had differentvalues or ranges than were supported in the DOE database, or they had questionable values orranges.

Each of these parameters was further reviewed by the Agency and dispositioned either by beingresolved and no longer in question, or by being included in the mandated PAVT with revisedvalues, ranges, or distributions determined by the Agency. The parameters and their dispositionsare discussed in Sections 3, 4, and 5 of this report and summarized in Tables 6.1 through 6.3. Some parameters were found to be no longer in question because they were either found to benot sensitive in the Agency’s sensitivity analysis, or because they were accepted after review ofadditional documentation provided by DOE or through the Agency’s studies. Other parameterswere no longer in question because they were eliminated from the PA model in response to anAgency-mandated change, or they were found to not have been used in the CCA version of thePA model. In the last case, if the parameter was considered important, alternate parameters wereidentified and changed in the PAVT to achieve the Agency’s objective in questioning theoriginal parameter.

After making the necessary adjustments to allow for model changes, a final list of 22 parametersthat were to be changed in the PAVT was developed and is presented in Table 6.4. This tablealso summarizes the parameter values, ranges, and distributions used in the PAVT as well as theoriginal values used in the CCA. The basis for selecting each parameter, value, range, anddistribution is presented in Sections 3, 4, and 5 of this report.

53

Table 6.1. Parameters Lacking Supporting Data

No. IDNo.



1 3245 BLOWOUT-CEMENT

Waste cementationstrength


Sensitive but removeddue to change in model

2 3246 BLOWOUT-PARTDIA

Waste particle diameter Required revisionusing expert elicitationprocess


3 198 DRZ_1-PRMX_LOG

Intrinsic permeability inX-direction in disturbedrock zone



4 2177 S_MB_139-DPHIMAX

Incremental increase inanhydrite porosity inMarker Bed 139


5 2180 S_MB_139-PF_DELTA

Incremental pressure forfull fracture development


6 586 S_MB_139-PI_DELTA

Fracture initiationpressure increment


7 2178 S_MB_139-KMAXLOG

Maximum permeabilityin altered anhydrite


8 3134 BH_OPEN-PRMX_LOG

Intrinsic permeability inX-direction in openborehole


9 2158 S_ANH_AB-DPHIMAX

Incremental increase inanhydrite porosity inbeds A and B


10 214 EXP_AREA-PRMX_LOG

Intrinsic permeability inX-direction inexperimental area


11 3473 BLOWOUT-THCK_CAS

Thickness of Castilebrine pocket for directbrine release


12 3456 BLOWOUT-RE_CAST

Radius of Castile brinepocket for direct brinerelease


13 3194 CASTILER-GRIDFLO

Index for selecting brinepocket volume

Removed in 4/25 letter Not used in CCA PAmodel - volume changedusing other parameters

55

Table 6.2. Parameters with Different Values or Ranges

No. IDNo.



1 3493 GLOBAL-PBRINE

Probability ofencountering pressurizedbrine



2 2254 BOREHOLE-TAUFAIL

Waste shear resistance Change required forPAVT in 4/25 letterand 6/6 note to docket


3 3184 BH_SAND-PRMX_LOG

Long term intrinsicborehole permeability inX-direction



4 2918 CASTILER-VOLUME

Castile brine pocketvolume

Removed in 4/25 letter Not used in CCA PAmodel - volume changedusing compressibilityand porosity adjustment

5 61 CASTILER-COMP_RCK

Castile brine pocket rockcompressibility


Not appropriatelyjustified; used to changebrine pocket volume

56

Table 6.3. Parameters with Questionable Values or Ranges

No. IDNo.



1 27 BOREHOLE-DOMEGA

Drill string angularvelocity



2 64 CASTILER-POROSITY

Castile brine pocketporosity


3 66 CASTILER-PRESSURE

Castile brine pocket porepressure


4 259 PAN_SEAL-PRMX_LOG

Intrinsic permeability ofpanel seal in X-direction


5 528 S_ANH_AB-POROSITY

Porosity of anhydritebeds A and B








8 651 WAS_AREA-ABSROUGH

Waste area absoluteroughness

Removed in 4/17 letter Not sensitive andacceptable after reviewof documentation

9 653 WAS_AREA-COMP_RCK

Waste area rockcompressibility


10 1992 WAS_AREA-DIRNCCHW

Bulk density of ironcontainers in CH waste

Removed in 4/17 letter Sensitive, butdocumentation providedand accepted afterreview

11 1993 WAS_AREA-DIRNCRHW

Bulk density of ironcontainers in RH waste


12 2040 WAS_AREA-DIRNCHW

Average density of iron-based material in CHwaste


13 2041 WAS_AREA-DCELLCHW

Average density ofcellulosics in CH waste


No. IDNo.



57

14 2274 WAS_AREA-DCELLRHW

Average density ofcellulosics in RH waste


15 2907 STEEL-CORRMCO2

Steel corrosion rate Change required forPAVT in 4/17 letter


16 3147 CONC_PLG-POROSITY

Borehole plug porosity Removed in 4/17 letter Not sensitive

17 3185 CONC_PLG-PRMX_LOG

Borehole plugpermeability in X-direction



18 3256 BLOWOUT-FGE

Gravity scaling factor Change required forPAVT in 4/17 letterbut removed

Removed due to changein model

19 3259 BLOWOUT-APORO

Waste permeability inCUTTINGS_S Model


Not used in CCA PAmodel -replaced withchanges to parameters663 WAS_AREA-PRMX_LOG and 2131REPOSIT-PRMX_LOG

20 3429 PHUMOX3-PHUMCIM



21 3471 BLOWOUT-MAXFLOW

Maximum period ofuncontrolled boreholeflow


22 3472 BLOWOUT-MINFLOW

Minimum period ofuncontrolled boreholeflow


23 3433 PHUMOX3-PHUMSIM



24 3470 GLOWOUT-GAS_MIN

DBR cutoff gas flow rate Removed in 4/25 letter Not sensitive

25 3317 PU-PROPMIC Microbial colloidproportionality constantfor plutonium



U(VI) solubility limit inCastile brine

Removed in 6/6 noteto docket

Sensitive, butdocumentation providedand accepted afterreview

No. IDNo.



58


U(VI) solubility limit inSalado brine

Removed in 6/6 noteto docket

Sensitive, butdocumentation providedand accepted afterreview

28a 3406 SOLMOD3-SOLCIM




28b 3402 SOLMOD3-SOLSIM




















33 3311 AM-PROPMIC Microbial colloidproportionality constantfor americium


34 3482 AM+3-MKD_AM

Matrix partitioncoefficient foramericium +3


Documentation providedand accepted afterreview

35 3480 PU+3-MKD_PU

Matrix partitioncoefficient for plutonium+3



36 3481 PU+4-MKD_PU

Matrix partitioncoefficient for plutonium+4



37 3479 U+4-MKD_U Matrix partitioncoefficient for uranium+4



38 3475 U+6-MKD_U Matrix partitioncoefficient for uranium+6



No. IDNo.



59

39 656 WAS_AREA-GRATMICH

Gas generation rate dueto microbial action underhumid conditions


40 657 WAS_AREA-GRATMICI

Gas generation rate dueto microbial action underinundated conditions


60

Table 6.4. Parameters Changed by EPA in Performance Assessment Verification Test



198 DRZ_1 -PRMX_LOG "

PAVTCCA

UniformConstant

-19.4-15.0

-12.5-15.0

-15.95-15.0

Log m2

Log m2

3184 BH_SAND -PRMX_LOG "

PAVTCCA

UniformUniform

-16.3-14.0

-11.0-11.0

-13.65-12.5

Log m2

Log m2

3185 CONC_PLG -PRMX_LOG "

PAVTCCA

UniformConstant

-19-16.3

-17-16.3

-17.3-16.3

Log m2

Log m2

663 WAS_AREA -PRMX_LOG*"

PAVTCCA

ConstantConstant

-12.6198-12.769

-12.6198-12.769

-12.6198-12.769

Log m2

Log m2

2131 REPOSIT -PRMX_LOG*"

PAVTCCA

ConstantConstant

-12.6198-12.769

-12.6198-12.769

-12.6198-12.769

Log m2

Log m2

2907 STEEL -CORRMCO2

PAVTCCA

UniformUniform

0.00.0

3.17 E-141.59 E-14

1.58 E-147.94 E-14

m/sm/s

61 CASTILER -COMP_RCK

PAVTCCA

TriangularTriangular

2.0 E-115.0 E-12

1.0 E-101.0 E-08

4.0 E-11-1.0 E-10-

Pa-1

Pa-1

8000 CASTILER -POR_BPKT j

PAVTCCA

Triangular- -

0.1848- -

0.9240- -

0.3696-- -

Dimension-less

3493 GLOBAL -PBRINE

PAVTCCA

UniformConstant

0.010.08

0.600.08

0.3050.08

Dimension-less

27 BOREHOLE -DOMEGA

PAVTCCA

CumulativeConstant

4.207.8

23.07.8

7.87.8

Radians/secRadians/sec

2254 BOREHOLE -TAUFAIL

PAVTCCA

LoguniformUniform

0.050.05

77.010.0

2.05.025

PaPa

3482 AM+3 -MKD_AM

PAVTCCA

LoguniformUniform

0.020.02

0.500.50

0.100.26

m3/kgm3/kg


LoguniformUniform

0.020.02

0.500.50

0.100.26

m3/kgm3/kg


LoguniformUniform

0.9000.900

20.020.0

4.24310.45

m3/kgm3/kg


LoguniformUniform

0.9000.900

20.020.0

4.24310.45

m3/kgm3/kg


LoguniformUniform

3.00 E-053.00 E-05

3.00 E-023.00 E-02

9.49 E-041.50 E-02

m3/kgm3/kg

3478 TH+4 - MKD_TH PAVTCCA

LoguniformUniform

0.9000.900

20.020.0

4.24310.45

m3/kgm3/kg



61


PAVTCCA

ConstantConstant

1.2 E-075.82 E-08

1.2 E-075.82 E-08

1.2 E-075.82 E-08



PAVTCCA

ConstantConstant

1.3 E-086.52 E-08

1.3 E-086.52 E-08

1.3 E-086.52 E-08



PAVTCCA

ConstantConstant

1.3 E-084.4 E-06

1.3 E-084.4 E-06

1.3 E-084.4 E-06



PAVTCCA

ConstantConstant

4.1 E-086.0 E-09

4.1 E-086.0 E-09

4.1 E-086.0 E-09



PAVTCCA

ConstantConstant

2.4 E-072.3 E-06

2.4 E-072.3 E-06

2.4 E-072.3 E-06



PAVTCCA

ConstantConstant

4.8 E-072.2 E-06

4.8 E-072.2 E-06

4.8 E-072.2 E-06


* These parameters replaced BLOWOUT - APORO, which is not used in a current PA code- This is the mode of the triangular distribution" Parameter similarly varied in Y- and Z-directionsj New parameter and number created by DOE for the PAVT to allow brine pocket volume to be varied

62

REFERENCES

Beauheim, R.L., W.R. Wawersik, and R.M. Roberts 1993. “Coupled Permeability andHydrofracture Tests to Assess the Waste-Containment Properties of FracturedAnhydrite,” Int. J. Rock Mech. Mineral Science and Geomechanics Abstracts, Vol. 30,pp. 1159-1163.

Beauheim, R.L., S.M. Howarth, P. Vaughn, S.W. Webb, and K.W. Larson 1994. IntegratedModeling and Experimental Programs to Predict Brine and Gas Flow at the WasteIsolation Plant. Report SAND94-0599C, Sandia National Laboratories, Albuquerque,New Mexico.

Boak, D.M., L. Dotson, R. Aguilar, D.W. Powers, and G. Newman 1996. Wellbore EnlargementInvestigation: Potential Analogs to the Waste Isolation Pilot Plant During Intrusion ofthe Repository. Report SAND96-2629, Sandia National Laboratories, Albuquerque, NewMexico.

CCA (DOE Carlsbad Area Office) 1996. Title 40 CFR Part 191, Compliance CertificationApplication for the Waste Isolation Pilot Plant. Report DOE/CAO-1996-2184, U.S.Department of Energy Waste Isolation Pilot Plant, Carlsbad Area Office, Carlsbad, NewMexico, October. Docket A-93-02, II-G-1.

Christensen, C.L., and T.O. Hunter 1980. The Bell Canyon Test Results. Report SAND80-2414C, Sandia National Laboratory, Albuquerque, New Mexico, November.

EPA 1998a, Technical Support Document for Section 194.24: WIPP Facility and WasteCharacterization, Section 4, Docket A-93-02, V-B-19.

EPA 1998b, Technical Support Document for Section 194.14: Assessment of Kds used in theCCA, Docket A-93-02, Item V-B-4.

EPA 1998c, Technical Support Document for Section 194.24: EPA’s Evaluation of DOE’sActinide Source Term, Docket A-93-02, Item V-B-17.

EPA 1998d, Technical Support Document for Section 194.23: Models and Computer Codes,Docket A-93-02, Item V-B-6.

DOE Carlsbad Area Office 1997. Expert Elicitation on WIPP Waste Particle Diameter SizeDistribution(s) During the 10,000-Year Regulatory Post-Closure Period, Draft. Preparedby the Carlsbad Area Office Technical Assistance Contractor for the U.S. Department ofEnergy Carlsbad Area Office, Carlsbad, New Mexico, May 12. Docket A-93-02, II-I-30.

Freeze, R.A., and J.A. Cherry 1979. Groundwater. Prentice-Hall Inc., Englewood Cliffs, NewJersey. Docket A-93-02, Reference #257.

63

Freeze, G.A., K.W. Larson, and P.B. Davies 1995. Coupled Multi-Phase Flow and ClosureAnalysis of Repository Response to Waste-Generated Gas at the Waste Isolation PilotPlant (WIPP). Report SAND93-1986, Sandia National Laboratories, Albuquerque, NewMexico. Docket A-93-02, Reference #261.

Kruger, M. 1997. Performance Assessment Parameters Values Identified in EPA Letters, DatedApril 17 and 25, 1997, to DOE. U.S. Environmental Protection Agency, Office ofRadiation and Indoor Air, Washington, D.C., June 6. Docket A-93-02, II-I-25 (April 17);II-I-27 (April 25).

Novak, C.F., R.C. Moore, and R.V. Bynum 1996. Prediction of Dissolved ActinideConcentrations in Concentrated Electrolyte Solutions: A Conceptual Model and ModelResults for the Waste Isolation Pilot Plant (WIPP). Proceedings of the 1996 InternationalConference on Deep Geological Disposal of Radioactive Waste, Canadian NuclearSociety, Toronto, Canada (ISSN No. 0-919784-44-5). Docket A-93-02, Reference #481.

Larson, K., R. Beauheim, and W. Weart 1997. Experimental Data and BRAGFLO FractureModel Parameter Values. Memorandum to Les Shephard, March 31, 1997, in letter fromGeorge E. Dials to Ramona Trovato. U.S. Department of Energy, Carlsbad Area Office,Carlsbad, New Mexico, April 15, 1997. Docket A-93-02, II-I-24, Comment #9.

Lenke, L.R., J.W. Berglund, and R.A. Cole 1996. Blowout Experiments Using Fine GrainedSilica Sands in an Axisymmetric Geometry. Report NMERI 1996/7/32250, New MexicoEngineering Research Institute, University of New Mexico,. Albuquerque, New Mexico,March. Docket A-93-02, Reference #396.

Parthenaides, E., and R.E. Paaswell 1970. Erodibility of Channels with Cohesive Boundary. J.Hydraulics Division, Proceedings, American Society of Civil Engineers, Vol. 96, pp.755-771, March. Docket A-93-02, Reference #502.

PR (Parameter Report) 1997. Technical Support Document for Section 194.23, ParameterReport. Prepared by TechLaw, Inc.,Lakewood, Colorado, for U.S. EnvironmentalProtection Agency, Office of Radiation and Indoor Air, Washington, D.C., August.Docket A-93-02, II-I-12.

Ross-Brown, D., J. Gibbons, D. Porter, and J. Schatz 1996. Final Waste Isolation Pilot PlantEngineered Systems Data Qualification Supplementary Peer Review Report. Prepared forU.S. Department of Energy, Carlsbad Area Office, Office of Regulatory Compliance,Carlsbad, New Mexico, December.

SA Report 1997. Technical Support Document for Section 194.23, Sensitivity Analysis Report.Prepared by TechLaw, Inc.,Lakewood , Colorado, for U.S. Environmental ProtectionAgency, Office of Radiation and Indoor Air, Washington, D.C., August. Docket A-93-02,II-I-13.

64

Simon D.B. and F. Senturk 1992. Sediment Transport Technology: Water and SedimentDynamics, in Sedimentation Engineering, Manual No. 54. American Society of CivilEngineers Water Resources Publication.

SNL 1992. Preliminary Performance Assessment for the Waste Isolation Pilot Plant, December1992, Volume 3: Model Parameters. Report SAND92-0700/3, Sandia NationalLaboratories, Albuquerque, New Mexico, December. Docket A-93-02, Reference #563.

U.S. Environemental Protection Agency, 1997b. Technical Support Document for Section194.14: Assessment of Kds used in the CCA. Prepared by TechLaw, Inc., Lakewood,Colorado, for U.S. Environmental Protection Agency, Office of Radiation and IndoorAir, Washington, D.C., September 2. Docket A-93-02, V-B-4.

Telander, M.R., and R.E. Westerman 1995. Hydrogen Generation by Metal Corrosion inSimulated Waste Isolation Pilot Plant Environments, Progress Report for the PeriodNovember 1989 through December 1992. Report SAND92-7347, Sandia NationalLaboratories, Albuquerque, New Mexico. Docket A-93-02, Reference #622.

Telander, M.R., and R.E. Westerman 1997. Hydrogen Generation by Metal Corrosion inSimulated Waste Isolation Pilot Plant Environments, Report SAND96-2538, SandiaNational Laboratories, Albuquerque, New Mexico.

The Earth Technology Corporation 1988. Final Report for Time Domain Electromagnetic(TDEM) Surveys at the WIPP Site. Report SAND87-7144, prepared by The EarthTechnology Corporation, Golden, Colorado, for Sandia National Laboratories,Albuquerque, New Mexico, June. Docket: A-93-02, Reference #229.

Trovato 1997A. Letter, E.R. Trovato (EPA) to A. Alm (DOE). Office of Air and Radiation, U.S.Environmental Protection Agency, Washington, D.C., 19 March. Docket A-93-02, II-I-17.

Trovato 1997B. Letter, E.R. Trovato (EPA) to G. Dials (DOE). Office of Air and Radiation, U.S.Environmental Protection Agency, Washington, D.C., 17 April. Docket A-93-02, II-I-25.

Trovato 1997C. Letter, E.R. Trovato (EPA) to G. Dials (DOE). Office of Air and Radiation, U.S.Environmental Protection Agency, Washington, D.C., 25 April. Docket A-93-02, II-I-27.

USBR 1987. Design of Small Dams. Third Edition, U.S. Bureau of Reclamation, Denver,Colorado.

Wilson, C., D. Porter, J. Gibbons, E. Oswald, G. Sjoblom, and F. Caporuscio 1997A. WasteIsolation Pilot Plant, Conceptual Models Second Supplementary Peer Review Report.U.S. Department of Energy Carlsbad Area Office, Carlsbad, New Mexico, January. Docket A-93-02, II-G-21.

65

Wilson, C., D. Porter, J. Gibbons, E. Oswald, G. Sjoblom, and F. Caporuscio 1997B. WasteIsolation Pilot Plant, Conceptual Models Third Supplementary Peer Review Report. U.S.Department of Energy Carlsbad Area Office, Carlsbad, New Mexico, April. Docket A-93-02, II-G-22.

WPO # 30603. Parameter Record Package for Salado Anhydrite Permeability in X-Direction bySusan Howarth, 4/30/96.

WPO # 30640. WIPP Shaft Seal System, BRAGFLO Parameters by L.D. Hurtado, 9/03/96.

WPO # 30819. Gas-Generation Parameters Required for BRAGFLO by Yifeng Wang and L.H.Brush, 2/07/96. Docket A-93-02, Reference #673.

WPO # 31084. Castile Brine Pocket Rock Compressibility Record Package by Geoff Freeze,2/07/96.

WPO # 31512. WIPP Data Entry Form 464, Drill String Angular Velocity by K. Rost, 1/10/97.

WPO # 32038. Revised DRZ Permeability Record Package by Rick Beauheim, 3/07/97.

WPO # 32328. Transuranic Waste Baseline Inventory Report (Revision 2). Report DOE/CAO-95-1121, U.S. Department of Energy, Carlsbad Area Office Technical AssistanceContractor, Carlsbad, New Mexico, December 1995. Docket A-93-02, II-G-1, VolumesIII to IV.

WPO # 34860. WIPP Data Entry Form 464, Effective Porosity for Marker Bed 139 by SusanHowarth, 2/14/96.

WPO # 34923. WIPP Data Entry Form 464, Humid Microbial Reaction Rate by Yifeng Wang,2/05/96.

WPO # 34928. WIPP Data Entry Form 464, Inundated Microbial Reaction Rate by YifengWang, 2/05/96.

WPO # 35162. Estimates of Gas Generation Parameters for the Long-Term PerformanceAssessment by Yifeng Wang and Larry Brush, 1/26/96.

WPO # 35181. An Adjustment for Using Steel Corrosion Rates in BRAFGLO to ReflectRepository Chemical Condition Changes due to Adding MgO as a Backfill by YifengWang and Larry Brush, 2/29/96.

WPO # 35197. Stability of Pu (VI), Np (VI), and U (VI) in Simulated WIPP Brine by D.T. Reed,D.G. Wygmans, and M.K. Richmann, 3/13/96.

66

WPO # 35695. Principal Investigator Documentation Package for Cuttings Release of SolidsCaused by Blowout by Barry M. Butcher, 9/05/96.

WPO # 35856. Parameter Record Package for Colloidal Actinide Source Term Parameters byH.W. Papenguth, 5//0/7/96.

WPO # 36488. Analysis of Uranium (VI) Solubility Data for WIPP Performance Assessment:Implementation of Analysis Plan AP-028 by D.E. Hobart and R.C. Moore, 7/31/96.Docket A-93-02, Reference #302.

WPO # 37148. Initial Pressure in the Castile Brine Reservoir by G. Freeze and K. Larson,3/20/96. Docket A-93-02, Reference #259

WPO # 37765. WIPP Performance Assessment User’s Manual for CUTTINGS_S Version 5.03,5/22/96. Docket A-93-02, II-G-3, Volume 4.

WPO # 37973.Castile Brine Reservoir Volume Revision by K. Larson and G. Freeze, 5/27/96.

WPO # 38241. Memorandum of Record, Justification for Assuming 1 psi for the CementationStrength Parameter in the CUTTINGS_S Model by Barry M. Butcher, 5/30/96

WPO # 38367. Porosity, Single-Phase Permeability, and Capillary Pressure Data fromPreliminary Laboratory Experiments on Selected Samples from Marker Bed 139 at theWaste Isolation Pilot Plant, Draft by S.M. Howarth and T. Christian-Frear. ReportSAND94-0472, Sandia National Laboratories, Albuquerque, New Mexico, 4/01/96.

WPO # 38801. Ranges and Probability Distributions for Dissolved Pu, Am, U, Th, and Np in theCulebra for the PA Calculations to Support the WIPP CCA. Memorandum from L.H.Brush to M.S. Tierney, 6/10/96. Docket A-93-02, Reference #93.

WPO # 39121.Portion of the Waste Panel Area, WIPP Site, that is Underlain by BrineReservoirs in the Castile Formation as Inferred from Transient Electromagnetic Method(TEM or TDEM) Surveys by D.J. Borns, 2/20/96.

WPO # 40199. Probability of Intercepting a Pressurized Brine Reservoir Under the WIPP byD.W. Powers, J.M. Sigda, and R.M. Holt, 7/10/96. Docket A-93-02, Reference #516.

WPO # 40434. Hardwired Values or Values that Differ from 1996 CCA Parameter Database byMary-Alena Martell, 8/20/96.

WPO # 40520. Analysis Package for the BRAGFLO Direct Release Calculations (Task 4) of thePerformance Assessment Analyses Supporting the CCA by Daniel M. Stoelzel and DarienG. O’Brien, 11./22.96. Docket A-93-02, II-G-05.

67

WPO # 40521. Analysis Package for Cuttings and Spalling Calculations (Tasks 5 and 6) of thePerformance Assessment Analyses Supporting the CCA by Jerry W. Berglund, 12/13/96. Docket A-93-02, II-G-06.

WPO # 41887. Treatment of Brine Reservoirs in CCA by Rick Beauheim, 8/22/96.

WPO # 42085. Brine Reservoirs in the Castile Formation, Waste Isolation Pilot Plant Project,Southeastern New Mexico, by R.S. Popielak, R.L. Beauheim, S.A. Black, W.E. Coons,C.T. Ellingson, and R.L. Olsen 1983. TME-3153, Vols 1 and 2. Westinghouse ElectricCorporation, Carlsbad, New Mexico. Docket: A-93-02, Reference #511.

WPO # 42248. Addendum to Colloidal Actinide Source Term Parameter Record Packages byH.W. Papenguth, 11/20/96.

WPO # 43672. Wellbore Enlargement Investigation: Potential Analogs to the Waste IsolationPilot Plant During Inadvertent Intrusion of the Repository by D.M. Boak, L. Dotson, R.Aguilar, D.W. Powers, and G. Newman. Report SAND96-2629, Sandia NationalLaboratories, Albuquerque, New Mexico, January 1997.

WPO # 44625. Actinide Stability/Solubility in Simulated WIPP Brines by D.T. Reed and D.G.Wygmans, Interim Report, 3/21/97.

WPO # 44699. Revisions to Castile Brine Reservoir Parameter Packages by Rick Beauheim,1/16/97.

WPO # 45115. U(VI) Solubility Calculation, Performance of Uranium (VI) SolubilityPredictions for EPA by R.V. Bynum and Y. Wang, 5/06/97.

WPO #46646. Estimate Waste Critical Shear Stress for WIPP PA Caving Model by YifengWang and K.W. Larson. Memorandum to M.S.Y. Chu and M.G. Marietta, SandiaNational Laboratory, Albuquerque, New Mexico, 6/27/97.

WPO #46936. Estimate WIPP Particle Sizes Based on Expert Elicitation Result: Revision 1 byYifeng Wang, Memorandum to M.S.Y. Chu and M.G. Marietta, Sandia NationalLaboratory, Albuquerque, New Mexico, 8/27/97.

docket no: a-93-02 v-b-14 2019/us epa... · docket no: a-93-02 v-b-14 technical support document...

Documents